Rozprawy doktorskie na temat „Evolutionary programming (computer science)”

The entire thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file; a non-technical public abstract appears in the public.pdf file. Title from PDF of title page (University of Missouri--Columbia, viewed on January 19, 2010). Thesis advisor: Dr. Tina Smilkstein. Includes bibliographical references.

A problem in Software Reuse (SR) is to find a software component appropriate to a given requirement. At present this is done by manual browsing through large libraries which is very time consuming and therefore expensi ve. Further to this, if the component is not the same as, but similar to a requirement, the component must be adapted to meet the requirements. This browsing and adaptation requires a skilled user who can comprehend library entries and foresee their application. It is expensive to train users and to produce these documented libraries. The specialised software design domain, chosen in this thesis, is that of Field Programmable Gate Arrays (FPGAs) programs. FPGAs are user programmable microchips that have many applications including encryption and control. This thesis is concerned with a specific technique for FGPA programming that uses Evolutionary Computing (EC) techniques to synthesize FPGA programs. Evolutionary Computing (EC) techniques are based on natural systems such as the life cycle of living organisms or the formation of crystalline structures. They can generate solutions to problems without the need for complete understanding of the problem. EC has been used to create software programs, and can be used as a knowledge-lean approach for generating libraries of software solutions. EC techniques produce solutions without documentation. To automate SR it has been shown that it is essential to understand the knowledge in the software library. In this thesis techniques for automatically documenting EC produced solutions are illustrated. It is also helpful to understand the principles at work in the reuse process. On examination of large collections of evolved programs it is shown that these programs contain reusable modules. Further to this, it is shown that by studying series of similar software components, principles of scale can be deduced. Case Based Reasoning (CBR) is a problem solving method that reuses old solutions to solve new problems and is an effective method of automatically reusing software libraries. These techniques enable automated creation, documentation and reuse of a software library. This thesis proposes that CBR is a feasible method for the reuse of EC designed FPGA programs. It is shown that EC synthesised FPGA programs can be documented, reused, and adapted to solve new problems, using automated CBR techniques.

Thesis (M.S. in Computer Science)--Naval Postgraduate School, December 1990.
Thesis Advisor(s): Shimeall, Timothy J. Second Reader: Barnes, Patrick D. "December 1990." Description based on title screen as viewed on March 29, 2010. DTIC Identifier(s): Computer Program Verification, Prototypes, Software Engineering, User Manuals. Author(s) subject terms: Software Testing, Software Prototyping, Rapid Prototyping, Reusable Components, Requirements-based Testing, Software Testing Tools. Includes bibliographical references (p. 276-281). Also available in print.

Thesis (M.S.)--Missouri University of Science and Technology, 2008.
Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed June 18, 2009) Includes bibliographical references (p. 62-64).

Natural gas companies are always trying to increase the accuracy of their forecasts. We introduce evolutionary programming as an approach to forecast natural gas demand more accurately. The created Evolutionary Programming Engine and Evolutionary Programming Ensemble Model use the current GasDay models, along with weather and historical flow to create an overall forecast for the amount of natural gas a company will need to supply to their customers on a given day. The existing ensemble model uses the GasDay component models and then tunes their individual forecasts and combines them to create an overall forecast.

The inputs into the Evolutionary Programming Engine and Evolutionary Programming Ensemble Model were determined based on currently used inputs and domain knowledge about what variables are important for natural gas forecasting. The ensemble model design is based on if-statements that allow different equations to be used on different days to create a more accurate forecast, given the expected weather conditions.

This approach is compared to what GasDay currently uses based on a series of error metrics and comparisons on different types of weather days and during different months. Three different operating areas are evaluated, and the results show that the created Evolutionary Programming Ensemble Model is capable of creating improved forecasts compared to the existing ensemble model, as measured by Root Mean Square Error (RMSE) and Standard Error (Std Error). However, the if-statements in the ensemble models were not able to produce individually reasonable forecasts, which could potentially cause errant forecasts if a different set of if-statements are true on a given day.

A hyper-heuristic is a search method or learning mechanism for selecting or generating heuristics to solve computational search problems. Researchers classify hyper-heuristics according to the source of feedback during learning: Online learning hyper-heuristics learn while solving a given instance of a problem; Offline learning hyper-heuristics learn from a set of training instances, a method that can generalise to unseen instances. Genetic programming (GP) can be considered a specialization of the more widely known genetic algorithms (GAs) where each individual is a computer program. GP automatically generates computer programs to solve specified tasks. It is a method of searching a space of computer programs. GP can be used as a kind of hyper-heuristic to be a learning algorithm when it uses some feedback from the search process. Our research mainly uses genetic programming as offline hyper-heuristic approach to automatically design various heuristics for evolutionary programming.

Thesis (M.S.)--Missouri University of Science and Technology, 2009.
Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed August 10, 2009) Includes bibliographical references (p. 49-51).

Thesis (Ph. D.)--George Mason University, 2007.
Title from PDF t.p. (viewed Jan. 22, 2008). Thesis director: Kenneth A. De Jong. Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Computer Science. Vita: p. 422. Includes bibliographical references (p. 413-421). Also available in print.

The aim of the main work in the thesis is to investigate Incremental Evolution methods for designing a suitable behavior arbitration mechanism for behavior-based (BB) robot controllers for autonomous mobile robots performing tasks of higher complexity. The challenge of designing effective controllers for autonomous mobile robots has been intensely studied for few decades. Control Theory studies the fundamental control principles of robotic systems. However, the technological progress allows, and the needs of advanced manufacturing, service, entertainment, educational, and mission tasks require features beyond the scope of the standard functionality and basic control. Artificial Intelligence has traditionally looked upon the problem of designing robotics systems from the high-level and top-down perspective: given a working robotic device, how can it be equipped with an intelligent controller. Later approaches advocated for better robustness, modifiability, and control due to a bottom-up layered incremental controller and robot building (Behavior-Based Robotics, BBR). Still, the complexity of programming such system often requires manual work of engineers. Automatic methods might lead to systems that perform task on demand without the need of expert robot programmer. In addition, a robot programmer cannot predict all the possible situations in the robotic applications. Automatic programming methods may provide flexibility and adaptability of the robotic products with respect to the task performed. One possible approach to automatic design of robot controllers is Evolutionary Robotics (ER). Most of the experiments performed in the field of ER have achieved successful learning of target task, while the tasks were of limited complexity. This work is a marriage of incremental idea from the BBR and automatic programming of controllers using ER. Incremental Evolution allows automatic programming of robots for more complex tasks by providing a gentle and easy-to understand support by expertknowledge — division of the target task into sub-tasks. We analyze different types of incrementality, devise new controller architecture, implement an original simulator compatible with hardware, and test it with various incremental evolution tasks for real robots. We build up our experimental field through studies of experimental and educational robotics systems, evolutionary design, distributed computation that provides the required processing power, and robotics applications. University research is tightly coupled with education. Combining the robotics research with educational applications is both a useful consequence as well as a way of satisfying the necessary condition of the need of underlying application domain where the research work can both reflect and base itself.

Thesis (M.S.)--Missouri University of Science and Technology, 2008.
Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed April 26, 2008) Includes bibliographical references (p. 65-68).

This Thesis expands the theoretical research done in the area of evolutionary algorithms. The (1+1)EA is a simple algorithm which allows to gain some insight in the behaviour of these randomized search heuristics. This work shows ways to possible improve on existing bounds. The general good runtime of the algorithm on linear functions is also proven for classes of quadratic functions. These classes are defined by the relative size of the quadratic and the linear weights. One proof of the paper looks at a worst case algorithm which always shows a worst case behaviour than many other functions. This algorithm is used as an upper bound for a lot of different classes.
Department of Computer Science

Advanced technologies, a world-wide trend to deregulation of power systems and environmental constraints have attracted increasing interest in the operation of electric power systems with multiple storage resources. Under the competitive pressure of the deregulation, new efficient solution techniques to adapt quickly to the changing marketplace are in demand. This thesis presents a new evolutionary method, Constructive Evolutionary Programming (CEP), for minimising the system operational cost of scheduling electric power systems with multiple storage resources. The method combines the advantages of Constructive Dynamic Programming and Evolutionary Programming. Instead of evolving the "primal" variables such as storage content releases and thermal generator outputs, CEP evolves the piecewise linear convex cost-to-go functions (i.e. the storage content value curves). The multi-stage problem of multi-storage power system scheduling is thus decomposed into many smaller one-stage subproblems with evolved cost-to-go functions. For each evolutionary individual, linear programming is used in the forward pass process to solve the dispatch subproblems and the total system operational cost over the scheduling period is assigned to its fitness. Case studies demonstrate that the proposed method is robust and efficient for multi-storage power systems, particularly large complex hydrothermal system with cascaded and pumped storages. Although the proposed method is in the early stage of development, relying on assumptions of piecewise linear convexity in a deterministic environment, methods for the incorporation of stochastic models, electrical network and nonlinear, non-convex and non-smooth models are discussed. In addition, a number of possible improvements are also outlined. Due to its simplicity but robustness and efficiency, there are potential research directions for the further development of this evolutionary optimisation method.

Traditionally, an evolutionary algorithm (EA) operates on a single population with no restrictions on possible mating pairs. Interesting changes to the behaviour of EAs emerge when the structure of the population is altered so that mating between individuals is restricted. Variants of EAs that use such populations are grouped into the field of spatially-structured EAs (SSEAs). Previous research into the behaviour of SSEAs has primarily focused on the impact space has on the selection pressure in the system. Selection pressure is usually characterised by takeover times and the ratio between the neighbourhood size and the overall dimension of space. While this research has given indications into where and when the use of an SSEA might be suitable, it does not provide a complete coverage of system behaviour in SSEAs. This thesis presents new research into areas of SSEA behaviour that have been left either unexplored or briefly touched upon in current EA literature. The behaviour of genetic drift in finite panmictic populations is well understood. This thesis attempts to characterise the behaviour of genetic drift in spatially-structured populations. First, an empirical investigation into genetic drift in two commonly encountered topologies, rings and torii, is performed. An observation is made that genetic drift in these two configurations of space is independent of the genetic structure of individuals and additive of the equivalent-sized panmictic population. In addition, localised areas of homogeneity present themselves within the structure purely as a result of drifting. A model based on the theory of random walks to absorbing boundaries is presented which accurately characterises the time to fixation through random genetic drift in ring topologies. A large volume of research has gone into developing niching methods for solving multimodal problems. Previously, these techniques have used panmictic populations. This thesis introduces the concept of localised niching, where the typically global niching methods are applied to the overlapping demes of a spatially structured population. Two implementations, local sharing and local clearing are presented and are shown to be frequently faster and more robust to parameter settings, and applicable to more problems than their panmictic counterparts. Current SSEAs typically use a single fitness function across the entire population. In the context of multimodal problems, this means each location in space attempts to discover all the optima. A preferable situation would be to use the inherent spatial properties of an SSEA to localise optimisation of peaks. This thesis adapts concepts from multiobjective optimisation with environmental gradients and applies them to multimodal problems. In addition to adapting to the fitness landscape, individuals evolve towards their preferred environmental conditions. This has the effect of separating individuals into regions that concentrate on different optima with the global fitness function. The thesis also gives insights into the expected number of individuals occupying each optima in the problem. The SSEAs and related models developed in this thesis are of interest to both researchers and end-users of evolutionary computation. From the end-user�s perspective, the developed SSEAs require less a priori knowledge of a given problem domain in order to operate effectively, so they can be more readily applied to difficult, poorly-defined problems. Also, the theoretical findings of this thesis provides a more complete understanding of evolution within spatially-structured populations, which is of interest not only to evolutionary computation practitioners, but also to researchers in the fields of population genetics and ecology.

Process Algebras are a Formal Modelling methodology which are an effective tool for defining models of complex systems, particularly those involving multiple interacting processes. However, describing such a model using Process Algebras requires expertise from both the modeller and the domain expert. Finding the correct model to describe a system can be difficult. Further more, even with the correct model, parameter tuning to allow model outputs to match experimental data can also be both difficult and time consuming. Evolutionary Algorithms provide effective methods for finding solutions to optimisation problems with large and noisy search spaces. Evolutionary Algorithms have been proven to be well suited to investigating parameter fitting problems in order to match known data or desired behaviour. It is proposed that Process Algebras and Evolutionary Algorithms have complementary strengths for developing models of complex systems. Evolutionary Algorithms require a precise and accurate fitness function to score and rank solutions. Process Algebras can be incorporated into the fitness function to provide this mathematical score. Presented in this work is the Evolving Process Algebra (EPA) framework, designed for the application of Evolutionary Algorithms (specifically Genetic Algorithms and Genetic Programming optimisation techniques) to models described in Process Algebra (specifically PEPA and Bio-PEPA) with the aim of evolving fitter models. The EPA framework is demonstrated using multiple complex systems. For PEPA this includes the dining philosophers resource allocation problem, the repressilator genetic circuit, the G-protein cellular signal regulators and two epidemiological problems: HIV and the measles virus. For Bio-PEPA the problems include a biochemical reactant-product system, a generic genetic network, a variant of the G-protein system and three epidemiological problems derived from the measles virus. Also presented is the EPA Utility Assistant program; a lightweight graphical user interface. This is designed to open the full functionality and parallelisation of the EPA framework to beginner or naive users. In addition, the assistant program aids in collating and graphing after experiments are completed.

Thesis (M.S.)--Missouri University of Science and Technology, 2010.
Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed July 19, 2010) Includes bibliographical references (p. 57-60).

Nature has created complex biological structures that exhibit intelligent behaviour through an evolutionary process. Thus, intelligence and evolution are intimately connected. This has inspired evolutionary computation (EC) that simulates the evolutionary process to develop powerful techniques such as genetic algorithms (GAs), genetic programming (GP), evolutionary strategies (ES) and evolutionary programming (EP) to solve real-world problems in learning, control, optimization and classification. GP discovers the relationship among data and expresses it as a LISP-S expression i.e., a computer program. Thus the goal of program discovery as a solution for a problem is addressed by GP in the framework of evolutionary computation. In this thesis, we address for the first time the problem of applying GP to mu1ticategory pattern classification. In supervised pattern classification, an input vector of m dimensions is mapped onto one of the n classes. It has a number of application areas such as remote sensing, medical diagnosis etc., A supervised classifier is developed by using a training set that contains representative samples of various classes present in the application. Supervised classification has been done earlier with maximum likelihood classifier: neural networks and fuzzy logic. The major considerations in applying GP to pattern classification are listed below: (i) GP-based techniques are data distribution-free i.e., no a priori knowledge is needed abut the statistical distribution of the data or no assumption such as normal distribution for data needs to be made as in MLC. (ii) GP can directly operate on the data in its original form. (iii) GP can detect the underlying but unknown relationship that mists among data and express it as a mathematical LISP S-expression. The generated LISP S-expressions can be directly used in the application environment. (iv) GP can either discover the most important discriminating features of a class during evolution or it requires minor post-processing of the LISP-S expression to discover the discriminant features. In a neural network, the knowledge learned by the neural network about the data distributions is embedded in the interconnection weights and it requires considerable amount of post-processing of the weights to understand the decision of the neural network. In 2-category pattern classification, a single GP expression is evolved as a discriminant function. The output of the GP expression can be +l for samples of one class and -1 for samples of the other class. When the GP paradigm is applied to an n-class problem, the following questions arise: Ql. As a typical GP expression returns a value (+l or -1) for a 2-class problem, how does one apply GP for the n-class pattern classification problem? Q2. What should be the fitness function during evolution of the GP expressions? Q3. How does the choice of a function set affect the performance of GP-based classification? Q4. How should training sets be created for evaluating fitness during the evolution of GP classifier expressions? Q5. How does one improve learning of the underlying data distributions in a GP framework? Q6. How should conflict resolution be handled before assigning a class to the input feature vector? Q7. How does GP compare with other classifiers for an n-class pattern classification problem? The research described here seeks to answer these questions. We show that GP can be applied to an n-category pattern classification problem by considering it as n 2-class problems. The suitability of this approach is demonstrated by considering a real-world problem based on remotely sensed satellite images and Fisher's Iris data set. In a 2-class problem, simple thresholding is sufficient for a discriminant function to divide the feature space into two regions. This means that one genetic programming classifier expression (GPCE) is sufficient to say whether or not the given input feature vector belongs to that class; i.e., the GP expression returns a value (+1 or -1). As the n-class problem is formulated as n 2-class problems, n GPCEs are evolved. Hence, n GPCE specific training sets are needed to evolve these n GPCEs. For the sake of illustration, consider a 5-class pat tern classification problem. Let n, be the number of samples that belong to class j, and N, be the number of samples that do not belong to class j, (j = 1,..., 5). Thus, N1=n2+n3+n4+n5 N2=n1+n3+n4+n5 N3=n1+n2+n4+n5 N4=n1+n2+n3+n5 N5=n1+n2+n3+n4 Thus, When the five class problem is formulated as five 2-class problems. we need five GPCEs as discriminant functions to resolve between n1 and N1, n2 and N2, n3 and N3, n4 and N4 and lastly n5 and N5. Each of these five 2-class problems is handled as a separate 2-class problem with simple thresholding. Thus, GPCE# l resolves between samples of class# l and the remaining n - 1 classes. A training set is needed to evaluate the fitness of GPCE during its evolution. If we directly create the training set, it leads to skewness (as n1 < N1). To overcome the skewness, an interleaved data format is proposed for the training set of a GPCE. For example, in the training set of GPCE# l, samples of class# l are placed alternately between samples of the remaining n - 1 classes. Thus, the interleaved data format is an artifact to create a balanced training set. Conventionally, all the samples of a training set are fed to evaluate the fitness of every member of the population in each generation. We call this "global" learning 3s GP tries to learn the entire training set at every stage of the evolution. We have introduced incremental learning to simplify the task of learning for the GP paradigm. A subset of the training set is fed and the size of the subset is gradually increased over time to cover the entire training data. The basic motivation for incremental learning is to improve learning during evolution as it is easier to learn a smaller task and then to progress from a smaller task to a bigger task. Experimental results are presented to show that the interleaved data format and incremental learning improve the performance of the GP classifier. We also show that the GPCEs evolved with an arithmetic function set are able to track variation in the input better than GPCEs evolved with function sets containing logical and nonlinear elements. Hence, we have used arithmetic function set, incremental learning, and interleaved data format to evolve GPCEs in our simulations. AS each GPCE is trained to recognize samples belonging to its own class and reject samples belonging to other classes a strength of association measure is associated with each GPCE to indicate the degree to which it can recognize samples belonging to its own class. The strength of association measures are used for assigning a class to an input feature vector. To reduce misclassification of samples, we also show how heuristic rules can be generated in the GP framework unlike in either MLC or the neural network classifier. We have also studied the scalability and generalizing ability of the GP classifier by varying the number of classes. We also analyse the performance of the GP classifier by considering the well-known Iris data set. We compare the performance of classification rules generated from the GP classifier with those generated from neural network classifier, (24.5 method and fuzzy classifier for the Iris data set. We show that the performance of GP is comparable to other classifiers for the Iris data set. We notice that the classification rules can be generated with very little post-processing and they are very similar to the rules generated from the neural network and C4.5 for the Iris data set. Incremental learning influences the number of generations available for GP to learn the data distribution of classes whose d is -1 in the interleaved data format. This is because the samples belonging to the true class (desired output d is +1) are alternately placed between samples belonging to other classes i.e., they are repeated to balance the training set in the interleaved data format. For example, in the evolution of GPCE for class# l, the fitness function can be fed initially with samples of class#:! and subsequently with the samples of class#3, class#4 and class#. So in the evaluation of the fitness function, the samples of class#kt5 will not be present when the samples of class#2 are present in the initial stages. However, in the later stages of evolution, when samples of class#5 are fed, the fitness function will utilize the samples of both class#2 and class#5. As learning in evolutionary computation is guided by the evaluation of the fitness function, GPCE# l gets lesser number of generations to learn how to reject data of class#5 as compared to the data of class#2. This is because the termination criterion (i.e., the maximum number of generations) is defined a priori. It is clear that there are (n-l)! Ways of ordering the samples of classes whose d is -1 in the interleaved data format. Hence a heuristic is presented to determine a possible order to feed data of different classes for the GPCEs evolved with incremental learning and interleaved data format. The heuristic computes an overlap index for each class based on its spatial spread and distribution of data in the region of overlap with respect to other classes in each feature. The heuristic determines the order in which classes whose desired output d is –1 should be placed in each GPCE-specific training set for the interleaved data format. This ensures that GP gets more number of generations to learn about the data distribution of a class with higher overlap index than a class with lower overlap index. The ability of the GP classifier to learn the data distributions depends upon the number of classes and the spatial spread of data. As the number of classes increases, the GP classifier finds it difficult to resolve between classes. So there is a need to partition the feature space and identify subspaces with reduced number of classes. The basic objective is to divide the feature space into subspaces and hence the data set that contains representative samples of n classes into subdata sets corresponding to the subspaces of the feature space, so that some of the subdata sets/spaces can have data belonging to only p classes (p < n). The GP classifier is then evolved independently for the subdata sets/spaces of the feature space. This results in localized learning as the GP classifier has to learn the data distribution in only a subspace of the feature space rather than in the entire feature space. By integrating the GP classifier with feature space partitioning (FSP), we improve classification accuracy due to localized learning. Although serial computers have increased steadily in their performance, the quest for parallel implementation of a given task has continued to be of interest in any computationally intensive task since parallel implementation leads to a faster execution than a serial implementation As fitness evaluation, selection strategy and population structures are used to evolve a solution in GP, there is scope for a parallel implementation of GP classifier. We have studied distributed GP and massively parallel GP for our approach to GP-based multicategory pattern classification. We present experimental results for distributed GP with Message Passing Interface on IBM SP2 to highlight the speedup that can be achieved over the serial implementation of GP. We also show how data parallelism can be used to further speed up fitness evaluation and hence the execution of the GP paradigm for multicategory pat tern classification. We conclude that GP can be applied to n-category pattern classification and its potential lies in its simplicity and scope for parallel implementation. The GP classifier developed in this thesis can be looked upon as an addition to the earlier statistical, neural and fuzzy approaches to multicategory pattern classification.

Nature has created complex biological structures that exhibit intelligent behaviour through an evolutionary process. Thus, intelligence and evolution are intimately connected. This has inspired evolutionary computation (EC) that simulates the evolutionary process to develop powerful techniques such as genetic algorithms (GAs), genetic programming (GP), evolutionary strategies (ES) and evolutionary programming (EP) to solve real-world problems in learning, control, optimization and classification. GP discovers the relationship among data and expresses it as a LISP-S expression i.e., a computer program. Thus the goal of program discovery as a solution for a problem is addressed by GP in the framework of evolutionary computation. In this thesis, we address for the first time the problem of applying GP to mu1ticategory pattern classification. In supervised pattern classification, an input vector of m dimensions is mapped onto one of the n classes. It has a number of application areas such as remote sensing, medical diagnosis etc., A supervised classifier is developed by using a training set that contains representative samples of various classes present in the application. Supervised classification has been done earlier with maximum likelihood classifier: neural networks and fuzzy logic. The major considerations in applying GP to pattern classification are listed below: (i) GP-based techniques are data distribution-free i.e., no a priori knowledge is needed abut the statistical distribution of the data or no assumption such as normal distribution for data needs to be made as in MLC. (ii) GP can directly operate on the data in its original form. (iii) GP can detect the underlying but unknown relationship that mists among data and express it as a mathematical LISP S-expression. The generated LISP S-expressions can be directly used in the application environment. (iv) GP can either discover the most important discriminating features of a class during evolution or it requires minor post-processing of the LISP-S expression to discover the discriminant features. In a neural network, the knowledge learned by the neural network about the data distributions is embedded in the interconnection weights and it requires considerable amount of post-processing of the weights to understand the decision of the neural network. In 2-category pattern classification, a single GP expression is evolved as a discriminant function. The output of the GP expression can be +l for samples of one class and -1 for samples of the other class. When the GP paradigm is applied to an n-class problem, the following questions arise: Ql. As a typical GP expression returns a value (+l or -1) for a 2-class problem, how does one apply GP for the n-class pattern classification problem? Q2. What should be the fitness function during evolution of the GP expressions? Q3. How does the choice of a function set affect the performance of GP-based classification? Q4. How should training sets be created for evaluating fitness during the evolution of GP classifier expressions? Q5. How does one improve learning of the underlying data distributions in a GP framework? Q6. How should conflict resolution be handled before assigning a class to the input feature vector? Q7. How does GP compare with other classifiers for an n-class pattern classification problem? The research described here seeks to answer these questions. We show that GP can be applied to an n-category pattern classification problem by considering it as n 2-class problems. The suitability of this approach is demonstrated by considering a real-world problem based on remotely sensed satellite images and Fisher's Iris data set. In a 2-class problem, simple thresholding is sufficient for a discriminant function to divide the feature space into two regions. This means that one genetic programming classifier expression (GPCE) is sufficient to say whether or not the given input feature vector belongs to that class; i.e., the GP expression returns a value (+1 or -1). As the n-class problem is formulated as n 2-class problems, n GPCEs are evolved. Hence, n GPCE specific training sets are needed to evolve these n GPCEs. For the sake of illustration, consider a 5-class pat tern classification problem. Let n, be the number of samples that belong to class j, and N, be the number of samples that do not belong to class j, (j = 1,..., 5). Thus, N1=n2+n3+n4+n5 N2=n1+n3+n4+n5 N3=n1+n2+n4+n5 N4=n1+n2+n3+n5 N5=n1+n2+n3+n4 Thus, When the five class problem is formulated as five 2-class problems. we need five GPCEs as discriminant functions to resolve between n1 and N1, n2 and N2, n3 and N3, n4 and N4 and lastly n5 and N5. Each of these five 2-class problems is handled as a separate 2-class problem with simple thresholding. Thus, GPCE# l resolves between samples of class# l and the remaining n - 1 classes. A training set is needed to evaluate the fitness of GPCE during its evolution. If we directly create the training set, it leads to skewness (as n1 < N1). To overcome the skewness, an interleaved data format is proposed for the training set of a GPCE. For example, in the training set of GPCE# l, samples of class# l are placed alternately between samples of the remaining n - 1 classes. Thus, the interleaved data format is an artifact to create a balanced training set. Conventionally, all the samples of a training set are fed to evaluate the fitness of every member of the population in each generation. We call this "global" learning 3s GP tries to learn the entire training set at every stage of the evolution. We have introduced incremental learning to simplify the task of learning for the GP paradigm. A subset of the training set is fed and the size of the subset is gradually increased over time to cover the entire training data. The basic motivation for incremental learning is to improve learning during evolution as it is easier to learn a smaller task and then to progress from a smaller task to a bigger task. Experimental results are presented to show that the interleaved data format and incremental learning improve the performance of the GP classifier. We also show that the GPCEs evolved with an arithmetic function set are able to track variation in the input better than GPCEs evolved with function sets containing logical and nonlinear elements. Hence, we have used arithmetic function set, incremental learning, and interleaved data format to evolve GPCEs in our simulations. AS each GPCE is trained to recognize samples belonging to its own class and reject samples belonging to other classes a strength of association measure is associated with each GPCE to indicate the degree to which it can recognize samples belonging to its own class. The strength of association measures are used for assigning a class to an input feature vector. To reduce misclassification of samples, we also show how heuristic rules can be generated in the GP framework unlike in either MLC or the neural network classifier. We have also studied the scalability and generalizing ability of the GP classifier by varying the number of classes. We also analyse the performance of the GP classifier by considering the well-known Iris data set. We compare the performance of classification rules generated from the GP classifier with those generated from neural network classifier, (24.5 method and fuzzy classifier for the Iris data set. We show that the performance of GP is comparable to other classifiers for the Iris data set. We notice that the classification rules can be generated with very little post-processing and they are very similar to the rules generated from the neural network and C4.5 for the Iris data set. Incremental learning influences the number of generations available for GP to learn the data distribution of classes whose d is -1 in the interleaved data format. This is because the samples belonging to the true class (desired output d is +1) are alternately placed between samples belonging to other classes i.e., they are repeated to balance the training set in the interleaved data format. For example, in the evolution of GPCE for class# l, the fitness function can be fed initially with samples of class#:! and subsequently with the samples of class#3, class#4 and class#. So in the evaluation of the fitness function, the samples of class#kt5 will not be present when the samples of class#2 are present in the initial stages. However, in the later stages of evolution, when samples of class#5 are fed, the fitness function will utilize the samples of both class#2 and class#5. As learning in evolutionary computation is guided by the evaluation of the fitness function, GPCE# l gets lesser number of generations to learn how to reject data of class#5 as compared to the data of class#2. This is because the termination criterion (i.e., the maximum number of generations) is defined a priori. It is clear that there are (n-l)! Ways of ordering the samples of classes whose d is -1 in the interleaved data format. Hence a heuristic is presented to determine a possible order to feed data of different classes for the GPCEs evolved with incremental learning and interleaved data format. The heuristic computes an overlap index for each class based on its spatial spread and distribution of data in the region of overlap with respect to other classes in each feature. The heuristic determines the order in which classes whose desired output d is –1 should be placed in each GPCE-specific training set for the interleaved data format. This ensures that GP gets more number of generations to learn about the data distribution of a class with higher overlap index than a class with lower overlap index. The ability of the GP classifier to learn the data distributions depends upon the number of classes and the spatial spread of data. As the number of classes increases, the GP classifier finds it difficult to resolve between classes. So there is a need to partition the feature space and identify subspaces with reduced number of classes. The basic objective is to divide the feature space into subspaces and hence the data set that contains representative samples of n classes into subdata sets corresponding to the subspaces of the feature space, so that some of the subdata sets/spaces can have data belonging to only p classes (p < n). The GP classifier is then evolved independently for the subdata sets/spaces of the feature space. This results in localized learning as the GP classifier has to learn the data distribution in only a subspace of the feature space rather than in the entire feature space. By integrating the GP classifier with feature space partitioning (FSP), we improve classification accuracy due to localized learning. Although serial computers have increased steadily in their performance, the quest for parallel implementation of a given task has continued to be of interest in any computationally intensive task since parallel implementation leads to a faster execution than a serial implementation As fitness evaluation, selection strategy and population structures are used to evolve a solution in GP, there is scope for a parallel implementation of GP classifier. We have studied distributed GP and massively parallel GP for our approach to GP-based multicategory pattern classification. We present experimental results for distributed GP with Message Passing Interface on IBM SP2 to highlight the speedup that can be achieved over the serial implementation of GP. We also show how data parallelism can be used to further speed up fitness evaluation and hence the execution of the GP paradigm for multicategory pat tern classification. We conclude that GP can be applied to n-category pattern classification and its potential lies in its simplicity and scope for parallel implementation. The GP classifier developed in this thesis can be looked upon as an addition to the earlier statistical, neural and fuzzy approaches to multicategory pattern classification.

This thesis was inspired by the difficulties of artificial evolutionary systems in finding elegant and well structured, regular solutions. That is that the solutions found are usually highly disorganized, poorly structured and exhibit limited re-use, resulting in bloat and other problems. This is also true of previous developmental evolutionary systems, where structural regularity emerges only by chance. We hypothesise that these problems might be ameliorated by incorporating repeated evaluations on increasingly difficult problems in the course of a developmental process. This thesis introduces a new technique for learning complex problems from a family of structured increasingly difficult problems, Evolutionary Developmental Evaluation (EDE). This approach appears to give more structured, scalable and regular solutions to such families of problems than previous methods. In addition, the thesis proposes some bio-inspired components that are required by developmental evolutionary systems to take full advantage of this approach. The key part of this is the developmental process, in combination with a varying fitness function evaluated at multiple stages of development, generates selective pressure toward generalisation. This also means that parsimony in structure is selected for without any direct parsimony pressure. As a result, the system encourages the emergence of modularity and structural regularity in solutions. In this thesis, a new genetic developmental system called Developmental Tree Adjoining Grammar Guided Genetic Programming (DTAG3P), is implemented, embodying the requirements above. It is tested on a range of benchmark problems. The results indicate that the method generates more regularly-structured solutions than the competing methods. As a result, the system is able to scale, at least on the problem classes tested, to very complex instances the system encourages the emergence of modularity and structural regularity in solutions. In this thesis, a new genetic developmental system called Developmental Tree Adjoining Grammar Guided Genetic Programming (DTAG3P), is implemented, embodying the requirements above. It is tested on a range of benchmark problems. The results indicate that the method generates more regularly-structured solutions than competing methods. As a result, the system is able to scale, at least on the problem classes tested, to very complex problem instances.

Thesis (Ph. D.)--Electrical and Computer Engineering, Georgia Institute of Technology, 2007.
Committee Chair: Vachtsevanos, George J.; Committee Co-Chair: Litt, Brian; Committee Member: Butera, Robert J.; Committee Member: Echauz, Javier; Committee Member: Howard, Ayanna M.; Committee Member: Williams, Douglas B.

For many real-world applications of evolutionary computation, the fitness function is obscured by random noise. This interferes with the evaluation and selection processes and adversely affects the performance of the algorithm. Noise can be effectively eliminated by averaging a large number of fitness samples for each candidate, but the number of samples used per candidate (the resampling rate) required to achieve this is usually prohibitively large and time-consuming. Hence there is a practical need for algorithms that handle noise without eliminating it. Moreover, the amount of noise (noise strength and distribution) may vary throughout the search space, further complicating matters. We study noisy problems for which the noise strength varies throughout the search space. Such problems have generally been ignored by previous work, which has instead generally focussed on the specific case where the noise strength is the same at all points in the search domain. However, this need not be the case, and indeed this assumption is false for many applications. For example, in games of chance such as Poker, some strategies may be more conservative than others and therefore less affected by the inherent noise of the game. This thesis makes three significant contributions in the field of noisy fitness functions: We present the concept of dynamic resampling. Dynamic resampling is a technique that varies the resampling rate based on the noise strength and fitness for each candidate individually. This technique is designed to exploit the variation in noise strength and fitness to yield a more efficient algorithm. We present several dynamic resampling algorithms and give results that show that dynamic resampling can perform significantly better than the standard resampling technique that is usually used by the optimisation community, and that dynamic resampling algorithms that vary their resampling rates based on both noise strength and fitness can perform better than algorithms that vary their resampling rate based on only one of the above. We study a specific class of noisy fitness functions for which we counterintuitively find that it is better to use a higher resampling rate in regions of lower noise strength, and vice versa. We investigate how the evolutionary search operates on such problems, explain why this is the case, and present a hypothesis (with supporting evidence) for classifying such problems. We present an adaptive engine that automatically tunes the noise compensation parameters of the search during the run, thereby eliminating the need for the user to choose these parameters ahead of time. This means that our techniques can be readily applied to real-world problems without requiring the user to have specialised domain knowledge of the problem that they wish to solve. These three major contributions present a significant addition to the body of knowledge for noisy fitness functions. Indeed, this thesis is the first work specifically to examine the implications of noise strength that varies throughout the search domain for a variety of noise landscapes, and thus starts to fill a large void in the literature on noisy fitness functions.

Thesis (Ph. D.)--Missouri University of Science and Technology, 2009.
Vita. The entire thesis text is included in file. Title from title screen of thesis/dissertation PDF file (viewed January 5, 2010) Includes bibliographical references (p. 152-159).

The field of neuroimaging has truly become data rich, and as such, novel analytical methods capable of gleaning meaningful information from large stores of imaging data are in high demand. Those methods that might also be applicable on the level of individual subjects, and thus potentially useful clinically, are of special interest. In this dissertation we introduce just such a method, called nonlinear functional mapping (NFM), and demonstrate its application in the analysis of resting state fMRI (functional Magnetic Resonance Imaging) from a 242-subject subset of the IMAGEN project, a European study of risk-taking behavior in adolescents that includes longitudinal phenotypic, behavioral, genetic, and neuroimaging data. Functional mapping employs a computational technique inspired by biological evolution to discover and mathematically characterize interactions among ROI (regions of interest), without making linear or univariate assumptions. Statistics of the resulting interaction relationships comport with recent independent work, constituting a preliminary cross-validation. Furthermore, nonlinear terms are ubiquitous in the models generated by NFM, suggesting that some of the interactions characterized here are not discoverable by standard linear methods of analysis. One such nonlinear interaction is discussed in the context of a direct comparison with a procedure involving pairwise correlation, designed to be an analogous linear version of functional mapping. Another such interaction suggests a novel distinction in brain function between drinking and non-drinking adolescents: a tighter coupling of ROI associated with emotion, reward, and interceptive processes such as thirst, among drinkers. Finally, we outline many improvements and extensions of the methodology to reduce computational expense, complement other analytical tools like graph-theoretic analysis, and possibly allow for voxel level functional mapping to eliminate the necessity of ROI selection.

This research presents a solution to the problem of tuning a PID controller for a nonlinear system. Many systems in industrial applications use a PID controller to control a plant or the process. Conventional PID controllers work in linear systems but are less effective when the plant or the process is nonlinear because PID controllers cannot adapt the gain parameters as needed. In this research we design a Nonlinear PID (NPID) controller using a fuzzy logic system based on the Mamdani type Fuzzy Inference System to control three different DC motor systems. This fuzzy system is responsible for adapting the gain parameters of a conventional PID controller. This fuzzy system's rule base was heuristically evolved using an Evolutionary Algorithm (Differential Evolution). Our results show that a NPID controller can restore a moderately or a heavily under-damped DC motor system under consideration to a desired behavior (slightly under-damped).

O problema da predição da estrutura de proteínas (PPEP) é bastante conhecido na bioinformática. A identificação da conformação nativa de uma proteína permite predizer a sua função no organismo. Este conhecimento também é útil no desenvolvimento de novos fármacos ou na compreensão do mecanismo de várias doenças. Várias técnicas tem sido propostas para resolver este problema. Porém, o alto custo envolvido levou ao surgimento de vários modelos que simplificam, em parte, as estruturas protéicas. No entanto, mesmo com os modelos mais simplificados, a complexidade do problema traz inúmeros desafios computacionais na busca da sua conformação nativa. Este trabalho utiliza o algoritmo evolucionário denominado Evolução Diferenciada (ED) para solucionar o PPEP, representando as proteínas com o modelo AB (toy model), em duas e três dimensões (2D e 3D). O trabalho apresenta a implementação de duas versões da ED, paralelizadas num ambiente de processo em cluster, com Message Passing Interface e arquitetura mestre-escravo. Para a configuração dos operadores do algoritmo de ED, foram realizados vários estudos com diferentes configurações para ambos os modelos, e análises estatísticas determinaram quais os melhores valores. Além disso, foram criados dois operadores especiais: dizimação e mutação espelhada. O primeiro poder ser considerado um operador genérico, que pode ser utilizado em qualquer problema; o segundo é específico para o problema em questão. Além do algoritmo de ED básico, também foi proposta uma versão auto-adaptável, em que alguns de seus parâmetros são atualizados no decorrer da evolução. Os experimentos realizados utilizaram 4 sequências de aminoácidos de benchmark geradas a partir da sequência de Fibonacci, contendo entre 13 e 55 aminoácidos. Os resultados dos algoritmos de ED paralelos foram comparados com os resultados obtidos em outros trabalhos. O algoritmo de ED é capaz de obter resultados excelentes, competitivos com os métodos especializados, apesar de não atingir o ótimo conhecido em algumas instâncias. Os resultados promissores obtidos nesse trabalho mostram que o algoritmo de ED é adequado para o problema. Em trabalhos futuros poderão ser estudados novos operadores especiais ou outras técnicas de inspiração biológica, buscando melhorar os resultados.
Protein structure prediction is a well-known problem in bioinformactis. Identifying protein native conformation makes it possible to predict its function within the organism. Knowing this also helps in the development of new medicines and in comprehending how some illnesses work and act. During the past year some techniques have been proposed to solve this problem, but its high cost made it necessary to build models that simplify the protein structures. However, even with the simplicity of these models identifying the protein native conformation remains a highly complex, computationally challenging problem. This paper uses an evolutionary algorithm known as Differential Evolution (DE) to solve the protein structure prediction problem. The model used to represent the protein structure is the Toy Model (also known as the AB Model) in both 2D and 3D. This work implements two versions of the ED algorithm using a parallel architecture (master-slave) based on Message Passing interface in a cluster. A large number of tests were executed to define the final configuration of the DE operators for both models. A new set of special operators were developed: explosion and mirror mutation. We can consider the first as generic, because it can be used in any problem. The second one is more specific because it requires previous knowledge of the problem. Of the two DE algorithm implemented, one is a basic DE algorithm and the second is a self-adaptive DE. All tests executed in this work used four benchmark amino acid sequences generated from the Fibonacci sequence. Each sequence has 13 to 55 amino acids. The results for both parallel DE algorithms using both 2D and 3D models were compared with other works. The DE algorithm achieved excellent results. It did not achieve the optimal known values for some sequences, but it was competitive with other specialized methods. Overall results encourage further research toward the use of knowledge-based operators and biologically inspired techniques to improve DE algorithm performance.

Evolutionary algorithms (EAs) have been successfully applied to many problems and applications. Their success comes from being general purpose, which means that the same EA can be used to solve different problems. Despite that, many factors can affect the behaviour and the performance of an EA and it has been proven that there isn't a particular EA which can solve efficiently any problem. This opens to the issue of understanding how different design choices can affect the performance of an EA and how to efficiently design and tune one. This thesis has two main objectives. On the one hand we will advance the theoretical understanding of evolutionary algorithms, particularly focusing on Genetic Programming and Parallel Evolutionary algorithms. We will do that trying to understand how different design choices affect the performance of the algorithms and providing rigorously proven bounds of the running time for different designs. This novel knowledge, built upon previous work on the theoretical foundation of EAs, will then help for the second objective of the thesis, which is to provide theory grounded design for Parallel Evolutionary Algorithms and Genetic Programming. This will consist in being inspired by the analysis of the algorithms to produce provably good algorithm designs.

Les problèmes d'optimisation sous contraintes non linéaires non convexes multi-objectifs en variables mixtes (discrètes et continues) apparaissent dans de nombreux domaines de l’ingénierie et notamment dans les applications de conception en mécaniques. Cette thèse vise à développer une nouvelle méthode pour résoudre ces problèmes d’optimisation. Notre proposition est une hybridation de l'algorithme multicritère « Branch-and-Bound » (MCBB) avec l’algorithme évolutionnaire de type NSGAII. L'approche proposée est en outre renforcée par de nouvelles stratégies de branchement conçues pour l’algorithme MCBB. Les contraintes du problème d’optimisation sont gérées à l'aide d'une nouvelle technique dédiée aux algorithmes évolutionnaires. Les performances de cette nouvelle approche sont évaluées et comparées à l’existant par une étude statistique sur un ensemble de problèmes tests. Les résultats montrent que les performances de notre algorithme sont compétitives face à l’algorithme NSGAII seul. Nous proposons deux applications de notre algorithme : les applications "Recherche de solutions faisables" et "Recherche de solutions optimales". Celles-ci sont appliquées sur un problème industriel réel d’un réducteur à engrenages à 3 étages formulé comme un problème bi-objectif. Dans ce problème des contraintes sont incluses pour satisfaire aux exigences de normes ISO sur le calcul de la capacité de charge des engrenages
Multi-objective mixed-integer non-convex non-linear constrained optimization problems that appears in several fields especially in mechanical applications. This thesis aims to develop a new method to solve such problems. Our proposal is a hybridization of the Multi-Criteria Branch-and-Bound (MCBB) algorithm with the Non-dominated Sorting Genetic Algorithm 2 (NSGAII). The proposed approach is furthermore enhanced by new branching strategies designed for MCBB. The constraints are handled using a new proposed constraint handling technique for evolutionary algorithms. Numerical experiments based on statistical assessment are done in this thesis to examine the performance of the new proposed approach. Results show the competitive performance of our algorithm among NSGAII. We propose two applications of our proposed approach: "Search Feasibility" and "Seek Optimality" applications. Both are applied on a real-world state of art 3 stages reducer problem which is formulated in this thesis to a bi-objective problem to meet the requirement of ISO standards on calculation of load capacity of gears

Many complex systems such as the Internet can be represented as networks, with vertices denoting the constituent components of the systems and edges denoting the patterns of interactions among the components. In this thesis, we are interested in how the structural properties of a network, such as its average degree, degree distribution, clustering, and homophily affect the processes that take place on it. In the first part of the thesis we focus on evolutionary game theory models for studying the evolution of cooperation in a population of predominantly selfish individuals. In the second part we turn our attention to an evolutionary model of disease dynamics and the impact of vaccination on the spread of infection. Throughout the thesis we use a network as an abstraction for a population, with vertices representing individuals in the population and edges specifying who can interact with whom. We analyze our models for a well-mixed population, i.e., an infinite population with random mixing, and compare the theoretical results with those obtained from computer simulations on model and empirical networks.

As the number of applications of computers controlling safety-critical operations increases, the need to ensure the safety and reliability of the software that controls those computers increases proportionally. Ultimately, such software properties are the result of appropriate design and implementation. However, certain characteristics of the language in which the software is written can have an impact on how that language facilitates both design and implementation, and how it encourages safety and reliability in the resulting software.
This paper explores the language characteristics that can impact the safety and reliability of the software produced. The goal is to provide a set of criteria that can be used for the selection of an appropriate language for real-time, safety-critical development. It proposes a set of characteristics that can affect the suitability of a language to such development, and compares a selection of common programming languages, including Ada, C, C++ and Java, against this framework.

The manual design of microarchitecture is labour-intensive and unlikely to generate good counterintuitive solutions. A designer, guided solely by intuition, is forced to manipulate many interdependent parameters by hand, introducing simplifying assumptions that artificially limit the potential of the architecture. The sheer number of possible configurations precludes an exhaustive exploration of the entire search space. Evolutionary computation is a domain-independent, highly scalable technique capable of producing innovative designs with reasonable computational cost.
This study presents a software framework for designing microarchitecture. As a case study, a genetic algorithm is used to automatically synthesize two-level indirect branch predictors. The evolved designs routinely deliver enhanced performance as compared to the equivalent, highly optimized structures created by hand. The primary drawback associated with this method is the excessive design complexity induced by unconstrained evolution. Several methods of incorporating simplicity of design into the evolutionary process are investigated. A unified fitness metric, combining both simplicity and performance, leads to the evolution of simple and effective designs.

The scaling problems which afflict attempts to optimise neural networks (NNs) with genetic algorithms (GAs) are disclosed. A novel GA-NN hybrid is introduced, based on the bumptree, a little-used connectionist model. As well as being computationally efficient, the bumptree is shown to be more amenable to genetic coding lthan other NN models. A hierarchical genetic coding scheme is developed for the bumptree and shown to have low redundancy, as well as being complete and closed with respect to the search space. When applied to optimising bumptree architectures for classification problems the GA discovers bumptrees which significantly out-perform those constructed using a standard algorithm. The fields of artificial life, control and robotics are identified as likely application areas for the evolutionary optimisation of NNs. An artificial life case-study is presented and discussed. Experiments are reported which show that the GA-bumptree is able to learn simulated pole balancing and car parking tasks using only limited environmental feedback. A simple modification of the fitness function allows the GA-bumptree to learn mappings which are multi-modal, such as robot arm inverse kinematics. The dynamics of the 'geographic speciation' selection model used by the GA-bumptree are investigated empirically and the convergence profile is introduced as an analytical tool. The relationships between the rate of genetic convergence and the phenomena of speciation, genetic drift and punctuated equilibrium arc discussed. The importance of genetic linkage to GA design is discussed and two new recombination operators arc introduced. The first, linkage mapped crossover (LMX) is shown to be a generalisation of existing crossover operators. LMX provides a new framework for incorporating prior knowledge into GAs. Its adaptive form, ALMX, is shown to be able to infer linkage relationships automatically during genetic search.

Knowledge Management has been the subject of increasing focus over the last several years. Literature and research on the topic has grown as companies and organizations have come to realize that success is often determined by one's ability to create, disseminate, and embody knowledge in products and services. This realization has led to increased interest in examining the ways in which knowledge can be effectively created, identified, codified, disseminated, and retained. The field of Knowledge Management has emerged to address this need.

One of the obstacles that Knowledge Management seeks to overcome is the natural tendency in people to hoard knowledge. People often withhold knowledge when they feel it provides them with a competitive advantage over others. Many traditional management incentives and team structures create and perpetuate competitive environments that encourage knowledge hoarding. Knowledge Management also seeks to find ways to reduce the impact of employee turnover. When an employee leaves a company or organization, the knowledge they possess often goes with them. This loss can potentially have a negative impact on the productivity and quality of the company or organization. Knowledge Management seeks to find ways to minimize loss of knowledge when an employee leaves a company or organization.

Pair programming is a practice that holds promise for overcoming some of the challenges faced by Knowledge Management. In pair programming, two programmers work side-by-side at one computer collaborating on the same design, algorithm, code, or test. The continual interaction between pair programmers would seem to provide an environment that promotes knowledge sharing, and collaborative knowledge discovery. Additionally, through pair rotation, in which pairs change partners fairly often, tacit knowledge might be spread more effectively through face-to-face communication than by documentation, databases, or other means.

This research examines pair programming in the realm of Knowledge Management, positing the following hypotheses:Pair programming reduces the tendency of people to hoard knowledge.Pair programming reduces the impact of employee turnover.Pair programming is an effective means of knowledge dissemination and knowledge retention that has a positive influence on the Knowledge Management practices of a company or organization.

These hypotheses are tested through the use of a survey of individuals in technology research, development, and service. Analysis of the survey results provided no conclusive evidence to either support or disprove the hypothesis that pair programming reduces the tendency of people to hoard knowledge. The results indicate support for the hypothesis that pair programming reduces the impact of employee turnover, although not statistically significant. And finally, the survey results indicate with statistical significance that pair programming is an effective means of knowledge dissemination and retention, with a positive influence on the Knowledge Management practices of a company or organization.