Dissertations / Theses on the topic 'Multiple Objective Genetic Algorithm (MOGA)'

Session ID: SA-G4-2
Joint 4th International Conference on Soft Computing and Intelligent Systems and 9th International Symposium on advanced Intelligent Systems, September 17-21, 2008, Nagoya University, Nagoya, Japan

La pyrolyse du bois est un processus crucial dans la science de la sécurité incendie car elle affecte la décomposition thermique et le comportement de combustion des matériaux. Le bois est un composite biopolymères (cellulose, hémicellulose et lignine) qui subit une pyrolyse complexe, produisant du charbon solide, du goudron et des gaz. Le processus de pyrolyse modifie également certaines caractéristiques importantes de l’échantillon (densité, conductivité thermique, capacité thermique, porosité, perméabilité, émissivité...) qui évoluent tout au long de la réaction de décomposition. La compréhension de ces transformations est cruciale pour la modélisation du comportement du feu des solides. Les évolutions des masses normalisées finales entre les expériences ATG et en cône calorimètre remettent en cause les modèles de taux de réaction solides existants. Les modèles actuels supposent souvent un ordre de réaction égal à 1, ce qui conduit à des inexactitudes lorsque l’ordre de réaction diffère de 1. Pour surmonter ces lacunes, un nouveau modèle basé sur la conversion, appelé ”Masse Initiale Virtuelle”, est proposé. Ce modèle est basé sur des données issues d’essais ATG. Il calcule la vitesse de chaque réaction dans le cas de mécanismes de pyrolyse complexes, avec de nombreuses réactions séquentielles et compétitives et a été implémenté en C++. Le code C++ de ce modèle est intégré avec l’outil DAKOTA pour permettre l’optimisation multi-objectif par algorithme génétique (MOGA) des paramètres cinétiques sur plusieurs vitesses de chauffage. Ce modèle de « Masse Initiale Virtuelle » est intégré dans la boîte à outils d’analyse des matériaux poreux basée sur OpenFOAM (PATO), un outil Open Source créé par la NASA. D’autres modèles de transferts de masse, de chaleur et de conservation des espèces en plus des propriétés des matériaux sont créés dans ce nouveau cadre. Un modèle informatique pour les réactions secondaires (réactions en phase gazeuse qui produisent du charbon secondaire) est implémenté dans PATO. Les simulations des essais en cône calorimètre sont effectuées dans des modèles 1D et 2D axisymétriques pour explorer l’influence des propriétés anisotropes du bois, en particulier l’orientation de ses fibres. La comparaison des modèles avec et sans réactions secondaires démontre le rôle de ces dernières dans la distribution de la chaleur et la production de charbon secondaire. Ce résultat explique la différence de masse finale observée expérimentalement entre les tests en ATG et en cône calorimètre. La comparaison des résultats expérimentaux et numériques montre la pertinence de cette approche pour simuler le comportement complexe de la pyrolyse du bois en mettant en évidence l’importance des voies de réaction, des réactions secondaires, du transfert de chaleur, du transfert de masse et des phénomènes d’interaction intra-pore
Pyrolysis of wood is a crucial process in fire safety science because it affects the thermal decomposition and combustion behavior of materials. Wood, a composite of biopolymeric components (cellulose, hemicellulose and lignin) undergoes complex pyrolysis to yield solid char, tar and gases as it thermally decomposes. The pyrolysis process also changes some important characteristics of the sample (density, thermal conductivity, heat capacity, porosity, permeability, emissivity...) that evolve throughout the reaction. Understanding these transformations is crucial for the correct modeling of fire behavior and material response under different thermal conditions. Different final normalized mass between TGA and cone calorimeter experiments challenge existing solid reaction rate models, according to experimental studies. Current models often assume a reaction order of 1, which oversimplifies the complexity of wood pyrolysis and leads to inaccuracies when the reaction order differs from 1. To overcome these shortcomings, a brand new conversion-based model, called ”Virtual Initial Mass”, is proposed. This model, based on TGA data, calculates the reaction rate for each reaction in complicated pyrolysis mechanisms. It supports mechanisms with numerous sequential and competitive reactions and has been implemented in C++. The C++ code for this model is integrated with the DAKOTA toolkit to perform multi objective genetic algorithm (MOGA) optimization of kinetic parameters for multiple heating rates. This ”Virtual Initial Mass” model is integrated in the Porous material Analysis Toolbox based on OpenFOAM (PATO) an Open Source tool distributed by NASA. Further mass transfer, heat transfer, species conservation models in addition to material properties are created within this new framework. A computational model for secondary reactions (gas-phase reactions that produce secondary char) is implemented in PATO. These secondary reactions solidify the sample and distribute heat back into the system. Simulations of cone calorimeter tests are performed in 1D and 2D axisymmetric models to explore the influence of anisotropic wood properties, particularly the orientation of wood fibers. Comparison of models with and without secondary reactions demonstrates their role in heat distribution and secondary char production and points out the experimentally observed difference in normalized mass between TGA and cone calorimeter tests. The model is verified by comparison with experimental results to show that it can simulate the complicated behavior of wood pyrolysis as well as emphasizes the importance of reaction pathways, secondary reactions, heat transfer, mass transfer and intra-pore interaction phenomena

This thesis introduces a multi-objective variation of the personnel assignment problem, by including additional hierarchical and team constraints, which put restrictions on possible matchings of the bipartite graph. Besides maximization of summation of weights that are assigned to the edges of the graph, these additional constraints are also treated as objectives which are subject to minimization. In this work, different genetic algorithm approaches to multi-objective optimization are considered to solve the problem. Weighted Sum &ndash
a classical approach, VEGA - a non-elitist multi-objective evolutionary algorithm, and SPEA &ndash
a popular elitist multi-objective evolutionary algorithm, are considered as means of solution to the problem, and their performances are compared with respect to a number of multi-objective optimization criteria.

The process developed herein uses a Multiple Objective Genetic Optimization (MOGO) algorithm. The optimization is implemented in ModelCenter (MC) from Phoenix Integration. It uses a genetic algorithm that searches the design space for optimal, feasible designs by considering three Measures of Performance (MOPs): Cost, Effectiveness, and Risk. The complete synthesis model is comprised of an input module, the three primary AUV synthesis modules, a constraint module, three objective modules, and a genetic algorithm. The effectiveness rating determined by the synthesis model is based on nine attributes identified in the US Navyâ s UUV Master Plan and four performance-based attributes calculated by the synthesis model. To solve multi-attribute decision problems the Analytical Hierarchy Process (AHP) is used. Once the MOGO has generated a final generation of optimal, feasible designs the decision-maker(s) can choose candidate designs for further analysis. A sample AUV Synthesis was performed and five candidate AUVs were analyzed.
Master of Science

Given the demand for mobility in our society, the cost of building additionalinfrastructures and the increasing concerns about the sustainability of the trafficsystem, traffic managers have to come up with new tools to optimize the trafficconditions within the existing infrastructure. This study considered to optimizethe durations of the green light phases in order to improve several criteria such asthe ability of the network to deal with important demands or the total pollutantemissions.     Because the modeling of the problem is difficult and computationally demanding,a stochastic micro-simulator called ’Simulation of Urban MObility’ (SUMO) has been used with a stochastic optimization process, namely a Genetic Algorithm (GA).     The research objective of the study was to create a computational frameworkbased on the integration of SUMO and a Multi-Objective Genetic-Algorithm (MOGA).The proposed framework was demonstrated on a medium-size network correspondingto a part of the town of Rouen, France. This network is composed of 11 intersections,168 traffic lights and 40 possible turning movements. The network is monitored with20 sensors, spread over the network. The MOGA considered in this study is basedon NSGA-II. Several aspects have been investigated during the course of this thesis.     An initial study shows that the proposed MOGA is successful in optimizing thesignal control strategies for a medium-sized network within a reasonable amount oftime.     A second study has been conducted to optimize the demand-related model ofSUMO in order to ensure that the behavior in the simulated environment is close tothe real one. The study shows that a hybrid algorithm composed of a gradient searchalgorithm combined with a GA achieved a satisfactory behavior2 for a medium-sizenetwork within a reasonable time.

Background: Multiple Objective Optimization problems(MOOPs) are common and evident in every field. Container port terminals are one of the fields in which MOOP occurs. In this research, we have taken a case in logistics management and modelled Multi-agent systems to solve the MOOP using Non-dominated Sorting Genetic Algorithm-II (NSGA-II). Objectives: The purpose of this study is to build AI-based models for solving a Multiple Objective Optimization Problem occurred in port terminals. At first, we develop a port agent with an objective function of maximizing throughput and a customer agent with an objective function of maximizing business profit. Then, we solve the problem using the single-objective optimization model and multi-objective optimization model. We then compare the results of both models to assess their performance. Methods: A literature review is conducted to choose the best algorithm among the existing algorithms, which were used previously in solving other Multiple Objective Optimization problems. An experiment is conducted to know how well the models performed to solve the problem so that all the participants are benefited simultaneously. Results: The results show that all three participants that are port, customer one and customer two have gained profits by solving the problem in multi-objective optimization model. Whereas in a single-objective optimization model, a single participant has achieved earnings at a time, leaving the rest of the participants either in loss or with minimal profits. Conclusion: We can conclude that multi-objective optimization model has performed better than the single-objective optimization model because of the impartial results among the participants.

Because of the increasing demand for the provision of energy worldwide and the numerous damages caused by a major use of fossil sources, the contribution of renewable energies has been increasing significantly in the global energy mix with the aim at moving towards a more sustainable development. In this context, this work aims at the development of a general methodology for designing PV systems based on ecodesign principles and taking into account simultaneously both techno-economic and environmental considerations. In order to evaluate the environmental performance of PV systems, an environmental assessment technique was used based on Life Cycle Assessment (LCA). The environmental model was successfully coupled with the design stage model of a PV grid-connected system (PVGCS). The PVGCS design model was then developed involving the estimation of solar radiation received in a specific geographic location, the calculation of the annual energy generated from the solar radiation received, the characteristics of the different components and the evaluation of the techno-economic criteria through Energy PayBack Time (EPBT) and PayBack Time (PBT). The performance model was then embedded in an outer multi-objective genetic algorithm optimization loop based on a variant of NSGA-II. A set of Pareto solutions was generated representing the optimal trade-off between the objectives considered in the analysis. A multi-variable statistical method (i.e., Principal Componet Analysis, PCA) was then applied to detect and omit redundant objectives that could be left out of the analysis without disturbing the main features of the solution space. Finally, a decision-making tool based on M-TOPSIS was used to select the alternative that provided a better compromise among all the objective functions that have been investigated. The results showed that while the PV modules based on c-Si have a better performance in energy generation, the environmental aspect is what makes them fall to the last positions. TF PV modules present the best trade-off in all scenarios under consideration. A special attention was paid to recycling process of PV module even if there is not yet enough information currently available for all the technologies evaluated. The main cause of this lack of information is the lifetime of PV modules. The data relative to the recycling processes for m-Si and CdTe PV technologies were introduced in the optimization procedure for ecodesign. By considering energy production and EPBT as optimization criteria into a bi-objective optimization cases, the importance of the benefits of PV modules end-of-life management was confirmed. An economic study of the recycling strategy must be investigated in order to have a more comprehensive view for decision making.

Microscopic traffic simulation has become an important tool to investigate traffic efficiency and road safety. In order to produce meaningful results, incorporated driver behaviour models need to be carefully calibrated to represent real world conditions. In addition to macroscopic relationships such as the speed-density diagram, they should also adequately represent the average risk of accidents occurring on the road. In this thesis, I present a two stage computationally feasible multi-objective calibration process. The first stage performs a parameter sensitivity analysis to select only parameters with considerable effect on the respective objective functions to keep the computational complexity of the calibration at a manageable level. The second stage employs a multi-objective genetic algorithm that produces a front of Pareto optimal solutions with respect to the objective functions. Compared to traditional methods which focus on only one objective while sacrificing accuracy of the other, my method achieves a high degree of realism for both traffic flow and average risk.
Mikroskopisk trafiksimulering har blivit ett viktigt verktyg för att undersöka trafik effektivitet och trafiksäkerhet. För att producera meningsfulla resultat måste inbyggda drivrutinsbeteendemodeller noggrant kalibreras för att representera verkliga förhållanden i världen. Förutom makroskopiska relationer, såsom hastighetsdensitetsdiagrammet, bör de också på ett adekvat sätt representera den genomsnittliga risken för olyckor som uppträder på vägen. I denna avhandling presenterar jag en tvåstegs beräkningsberättigbar mångsidig kalibreringsprocess. Det första steget utför en parameterkänslighetsanalysför att bara välja parametrar med stor effekt på respektive objektiv funktioner för att hålla kalibrerings komplexiteten på en hanterbar nivå. Det andra steget använder en mångriktig genetisk algoritm som ger framsidan av Pareto optimala lösningar med hänsyn till objektivfunktionerna. Jämfört med traditionella metoder som fokuserar på endast ett mål, samtidigt som man offrar den andra, ger min metod en hög grad av realism för både trafikflöde och genomsnittlig risk.

Les fédérations de nuages informatiques peuvent être considérées comme une avancée majeure dans l’informatique en nuage, en particulier dans le domaine médical. En effet, le partage de données médicales améliorerait la qualité des soins. La fédération de ressources permettrait d'accéder à toutes les informations, même sur une personne mobile, avec des données hospitalières distribuées sur plusieurs sites. En outre, cela permettrait d’envisager de plus grands volumes de données sur plus de patients et ainsi de fournir des statistiques plus fines. Les données médicales sont généralement conformes à la norme DICOM (Digital Imaging and Communications in Medicine). Les fichiers DICOM peuvent être stockés sur différentes plates-formes, telles qu’Amazon, Microsoft, Google Cloud, etc. La gestion des fichiers, y compris le partage et le traitement, sur ces plates-formes, suit un modèle de paiement à l’utilisation, selon des modèles de prix distincts et en s’appuyant sur divers systèmes de gestion de données (systèmes de gestion de données relationnelles ou SGBD ou systèmes NoSQL). En outre, les données DICOM peuvent être structurées en lignes ou colonnes ou selon une approche hybride (ligne-colonne). En conséquence, la gestion des données médicales dans des fédérations de nuages soulève des problèmes d’optimisation multi-objectifs (MOOP - Multi-Objective Optimization Problems) pour (1) le traitement des requêtes et (2) le stockage des données, selon les préférences des utilisateurs, telles que le temps de réponse, le coût monétaire, la qualités, etc. Ces problèmes sont complexes à traiter en raison de la variabilité de l’environnement (liée à la virtualisation, aux communications à grande échelle, etc.). Pour résoudre ces problèmes, nous proposons MIDAS (MedIcal system on clouD federAtionS), un système médical sur les fédérations de groupes. Premièrement, MIDAS étend IReS, une plate-forme open source pour la gestion de flux de travaux d’analyse sur des environnements avec différents systèmes de gestion de bases de données. Deuxièmement, nous proposons un algorithme d’estimation des valeurs de coût dans une fédération de nuages, appelé Algorithme de régression %multiple linéaire dynamique (DREAM). Cette approche permet de s’adapter à la variabilité de l'environnement en modifiant la taille des données à des fins de formation et de test, et d'éviter d'utiliser des informations expirées sur les systèmes. Troisièmement, l’algorithme génétique de tri non dominé à base de grilles (NSGA-G) est proposé pour résoudre des problèmes d’optimisation multi-crtières en présence d’espaces de candidats de grande taille. NSGA-G vise à trouver une solution optimale approximative, tout en améliorant la qualité du font de Pareto. En plus du traitement des requêtes, nous proposons d'utiliser NSGA-G pour trouver une solution optimale approximative à la configuration de données DICOM. Nous fournissons des évaluations expérimentales pour valider DREAM, NSGA-G avec divers problèmes de test et jeux de données. DREAM est comparé à d'autres algorithmes d'apprentissage automatique en fournissant des coûts estimés précis. La qualité de la NSGA-G est comparée à celle des autres algorithmes NSGA présentant de nombreux problèmes dans le cadre du MOEA. Un jeu de données DICOM est également expérimenté avec NSGA-G pour trouver des solutions optimales. Les résultats expérimentaux montrent les qualités de nos solutions en termes d'estimation et d'optimisation de problèmes multi-objectifs dans une fédération de nuages
Cloud federations can be seen as major progress in cloud computing, in particular in the medical domain. Indeed, sharing medical data would improve healthcare. Federating resources makes it possible to access any information even on a mobile person with distributed hospital data on several sites. Besides, it enables us to consider larger volumes of data on more patients and thus provide finer statistics. Medical data usually conform to the Digital Imaging and Communications in Medicine (DICOM) standard. DICOM files can be stored on different platforms, such as Amazon, Microsoft, Google Cloud, etc. The management of the files, including sharing and processing, on such platforms, follows the pay-as-you-go model, according to distinct pricing models and relying on various systems (Relational Data Management Systems or DBMSs or NoSQL systems). In addition, DICOM data can be structured following traditional (row or column) or hybrid (row-column) data storages. As a consequence, medical data management in cloud federations raises Multi-Objective Optimization Problems (MOOPs) for (1) query processing and (2) data storage, according to users preferences, related to various measures, such as response time, monetary cost, qualities, etc. These problems are complex to address because of heterogeneous database engines, the variability (due to virtualization, large-scale communications, etc.) and high computational complexity of a cloud federation. To solve these problems, we propose a MedIcal system on clouD federAtionS (MIDAS). First, MIDAS extends IReS, an open source platform for complex analytics workflows executed over multi-engine environments, to solve MOOP in the heterogeneous database engines. Second, we propose an algorithm for estimating of cost values in a cloud environment, called Dynamic REgression AlgorithM (DREAM). This approach adapts the variability of cloud environment by changing the size of data for training and testing process to avoid using the expire information of systems. Third, Non-dominated Sorting Genetic Algorithm based ob Grid partitioning (NSGA-G) is proposed to solve the problem of MOOP is that the candidate space is large. NSGA-G aims to find an approximate optimal solution, while improving the quality of the optimal Pareto set of MOOP. In addition to query processing, we propose to use NSGA-G to find an approximate optimal solution for DICOM data configuration. We provide experimental evaluations to validate DREAM, NSGA-G with various test problem and dataset. DREAM is compared with other machine learning algorithms in providing accurate estimated costs. The quality of NSGA-G is compared to other NSGAs with many problems in MOEA framework. The DICOM dataset is also experimented with NSGA-G to find optimal solutions. Experimental results show the good qualities of our solutions in estimating and optimizing Multi-Objective Problem in a cloud federation

Mathematical modeling and simulations of a submerged renewable storage system integrated to a wind farm, chosen based on the geographical and wind conditions at the Baltic Sea, gives insight on the feasibility of the submerged renewable storage and an approximation of the payback period and profits that could be generated. Genetic Algorithms were used to obtain the optimal number of spheres for a certain depth, based on 2 objective functions I.e. Minimum Life Cycle Cost (LCC) and maximum reduction in wind curtailment. The new arrangement concept shows that the Initial Capital Cost (ICC) could be decreased by 25% to 60% depending upon the number of sphere employed. Based on the inputs considered in the study, the results prove that the submerged renewable storage system would be feasible, and the profits ranging from 15 Million Euro to 29 Million Euro can be achieved at the chosen location, towards the Swedish side of the Baltic sea. Although, in a real life scenario it is assumed that only up to half of the profits obtained in the results would be achievable. The results also show that, the Pump/Turbine with a high turbine efficiency and lower pump efficiency, generated better profits, compared to a Pump/Turbine running with a higher pump efficiency and lower turbine efficiency. An attempt to increase the round-trip efficiency by adding a multi stage submersible pump, resulted in additional ICC and LCC, which saw a decrease in profits.

L’objectif de ce travail est de développer un cadre méthodologique et générique d’éco-conception de procédés chimiques couplant des outils de modélisation et de simulation traditionnels de procédés (HYSYS, COCO, ProSimPlus et Ariane), d’Analyse du Cycle de Vie (ACV), d’optimisation multiobjectif basée sur des Algorithmes Génétiques et enfin des outils d’aide à la décision multicritère (ELECTRE, PROMETHEE, M-TOPSIS). Il s’agit de généraliser, d’automatiser et d’optimiser l’évaluation des impacts environnementaux au stade préliminaire de la conception d’un procédé chimique. L’approche comprend trois étapes principales. Les deux premières correspondent d’une part aux phases d’analyse de l’inventaire par calcul des bilans de matière et d’énergie et d’autre part à l’évaluation environnementale par ACV. Le problème du manque d’information ou de l’imprécision dans les bases de données classiques en ACV pour la production d’énergie notamment sous forme de vapeur largement utilisée dans les procédés a reçu une attention particulière. Une solution proposée consiste à utiliser un simulateur de procédés de production d’utilités (Ariane, ProSim SA) pour contribuer à alimenter la base de données environnementale en tenant compte de variations sur les conditions opératoires ou sur les technologies utilisées. Des sous-modules « énergie » sont ainsi proposés pour calculer les émissions relatives aux impacts liés à l’utilisation de l’énergie dans les procédés. La troisième étape réalise l’interaction entre les deux premières phases et l’optimisation multi-objectif qui met en jeu des critères économiques et environnementaux. Elle conduit à des solutions de compromis le long du front de Pareto à partir desquelles les meilleures sont choisies à l’aide de méthodes d’aide à la décision. L’approche est appliquée à des procédés de production continus : production de benzène par hydrodéalkylation du toluène HDA et production de biodiesel à partir d’huiles végétales. Une stratégie à plusieurs niveaux est mise en oeuvre pour l'analyse de l'optimisation multi-objectif. Elle est utilisée dans les deux cas d'étude afin d'analyser les comportements antagonistes des critères
The objective of this work is to propose an integrated and generic framework for eco-design coupling traditional modelling and flowsheeting simulation tools (HYSYS, COCO, ProSimPlus and Ariane), Life Cycle Assessment, multi-objective optimization based on Genetic Algorithms and multiple criteria decision-making methods MCDM (Multiple Choice Decision Making, such as ELECTRE, PROMETHEE, M-TOPSIS) that generalizes, automates and optimizes the evaluation of the environmental criteria at earlier design stage. The approach consists of three main stages. The first two steps correspond respectively to process inventory analysis based on mass and energy balances and impact assessment phases of LCA methodology. Specific attention is paid to the main issues that can be encountered with database and impact assessment i.e. incomplete or missing information, or approximate information that does not match exactly the real situation that may introduce a bias in the environmental impact estimation. A process simulation tool dedicated to production utilities, Ariane, ProSim SA is used to fill environmental database gap, by the design of specific energy sub modules, so that the life cycle energy related emissions for any given process can be computed. The third stage of the methodology is based on the interaction of the previous steps with process simulation for environmental impact assessment and cost estimation through a computational framework. The use of multi-objective optimization methods generally leads to a set of efficient solutions, the so-called Pareto front. The next step consists in identifying the best ones through MCDM methods. The approach is applied to two processes operating in continuous mode. The capabilities of the methodology are highlighted through these case studies (benzene production by HDA process and biodiesel production from vegetable oils). A multi-level assessment for multi-objective optimization is implemented for both cases, the explored pathways depending on the analysis and antagonist behaviour of the criteria

Cette thèse est dédiée à l’optimisation de composants et de systèmes énergétiques avec une application dans le secteur résidentiel. La méthodologie développée est composée d’une analyse de sensibilité, d’une optimisation multi-objectif et d’une aide à la décision multicritère pour sélectionner le meilleur compromis. Tout d’abord, une optimisation d’un système de production d’eau chaude sanitaire est mise en œuvre numériquement et s’appuie sur un banc expérimental du laboratoire IUSTI. Cette étude a pour but d’optimiser, en fonction de différents profils de puisage, les performances d’un dispositif basé sur une pompe à chaleur en améliorant sa régulation. Ensuite, la procédure est appliquée à l’optimisation paramétrique d’un échangeur air-sol (EAS). Le système utilise les ressources géothermiques pour chauffer ou rafraîchir l’air d’un bâtiment par ventilation. Le modèle de l’échangeur air-sol a été validé expérimentalement avec une plateforme géothermique existante à Strasbourg. Un système qui couple un EAS, une ventilation double flux et une pompe à chaleur est également étudié. Le dimensionnement optimal permet d’obtenir un système qui demeure à la fois rentable, autonome et performant pour les différents climats considérés. Enfin, le processus est appliqué à l’optimisation topologique des échangeurs de chaleur. La procédure identifie l’ensemble des topologies qui possède un bon compromis entre les pertes de charge et les transferts thermiques. La méthodologie d'aide à la décision sélectionne la topologie finale qui permet d’avoir une répartition optimisée d’éléments solides afin d’obtenir le meilleur compromis entre ces objectifs
This thesis is dedicated to the optimization of components and energy systems with an application in the residential sector. The methodology developed is composed of a sensitivity analysis, a multi-objective optimization and a multi-criteria decision-making aid to select the best compromise.First of all, an optimization of a domestic hot water production system is implemented numerically and is based on an experimental set-up in the IUSTI laboratory. The aim of this study is to optimize the performance of a heat pump-based system by improving its regulation according to different drawing profiles. Then, the procedure is applied to the parametric optimization of an earth-air heat exchanger (EAHE). The system uses geothermal resources to preheat or cool the air in a building by ventilation. The model of the earth-air heat exchanger has been experimentally validated with an existing geothermal platform at Strasbourg. A system that combines an EAHE, a double flow ventilation and a heat pump is also studied. Optimal sizing makes it possible to obtain a system that is profitable, autonomous and efficient for the different climates considered. Finally, the process is applied to the topological optimization of heat exchangers. The procedure identifies the set of topologies that has a good compromise between pressure drops and heat transfer. The decision aid methodology selects the final topology that allows to have an optimized distribution of solid elements in order to obtain the best compromise between these objectives

碩士
中原大學
資訊管理研究所
96
This study proposes a new genetic evaluation method to improve the non-dominate sorting genetic algorithm-II (NSGA-II), which is a well-known algorithm for finding the Pareto-optimal set of multi-objective optimization problems. An evaluative-NSGA-II (E-NSGA-II) proposed in this thesis is a modified version of NSGA-II in which an evaluative crossover incorporates is cooperated to retain superior schema patterns in each chromosome to enhance the solution performance. The experiment results have been compared with eight existing algorithms on thirteen benchmark multi-objective problems, which include night unconstrained problems and four constrained problems. The results indicate that E-NSGA-II can find Pareto-optimal solutions in continuous test cases and is an effective method for solving multi-objective problems. As a whole, E-NSGA-II can achieve better convergence ability and great diversity quality than other algorithms.

碩士
國立臺灣科技大學
工業管理系
105
Solar energy is a potential project because it not only protects the environment but also reserves the power for people to use in their daily life such as heating or lighting. This study focuses on the natural sunlight saving system named solar concentrator layout. In our study, we aim to bring the optimal profit for the firm when implementing the solar layout as well as helping a house get as much sunlight efficiency as possible for their using. We also consider some factors such as light reflection and light transmission loss to make the model more reliable. As for the economic scale, some constraints are added to make our study close to reality, such as the thickness of concentrator or the number of exits where a sunbeam is delivered to the main panel to enable energy transmission. To obtain a high brightness for the house, the firms would harm their profit. This study makes the balance between the conflict objectives to get a compromised solution. Finally, parallel-computing based genetic algorithm is introduced to accelerate the solution quality and speed. To summarize, the result of our study will be the best strategy for the light efficiency to supply people and the profits that the firm earns for the job.

碩士
國立高雄應用科技大學
金融資訊研究所
99
Multi-objective optimization is to simultaneously optimize two or more conflict objectives to certain constraints. Because the solution space of multi-objective optimization is often a non-convex or discontinues shape, the conventional evaluation methods are hard to find an efficient frontier efficiently. The multi-objective genetic algorithm (MOGA) is a state-of-the-art nonlinear optimization methodology that applies weight-sum method or Pareto-based ranking schemes. These well-known MOGA methods, such as Non-dominated Sorting Genetic Algorithm-II (NSGA-II) and Strength Pareto Evolutionary Approach 2 (SPEA-2), maintains diversity solution set in the optimization process. However, these MOGA based co-evolution mechanisms, such as MOGAs with sexual selection, are presented to maintain more aggressive solution set. In this study, we introduce an improved MOGA, the Strengthen Dominant Species Genetic Algorithm (SDSGA) that proposed an enhanced selection mechanism with crowding estimation technique to extract more dominated species. The empirical results indicate that SDSGA outperforms in three and more objectives problems.

Given the demand for mobility in our society, the cost of building additionalinfrastructures and the increasing concerns about the sustainability of the trafficsystem, traffic managers have to come up with new tools to optimize the trafficconditions within the existing infrastructure. This study considered to optimizethe durations of the green light phases in order to improve several criteria such asthe ability of the network to deal with important demands or the total pollutantemissions.     Because the modeling of the problem is difficult and computationally demanding,a stochastic micro-simulator called ’Simulation of Urban MObility’ (SUMO) hasbeen used with a stochastic optimization process, namely a Genetic Algorithm (GA).     The research objective of the study was to create a computational frameworkbased on the integration of SUMO and a Multi-Objective Genetic-Algorithm (MOGA).The proposed framework was demonstrated on a medium-size network correspondingto a part of the town of Rouen, France. This network is composed of 11 intersections, 168 traffic lights and 40 possible turning movements. The network is monitored with20 sensors, spread over the network. The MOGA considered in this study is basedon NSGA-II. Several aspects have been investigated during the course of this thesis.     An initial study shows that the proposed MOGA is successful in optimizing the signal control strategies for a medium-sized network within a reasonable amount oftime.     A second study has been conducted to optimize the demand-related model ofSUMO in order to ensure that the behavior in the simulated environment is close tothe real one. The study shows that a hybrid algorithm composed of a gradient searchalgorithm combined with a GA achieved a satisfactory behavior for a medium-sizenetwork within a reasonable time. The demand is defined as the number of cars

碩士
國立臺北科技大學
工業工程與管理系碩士班
102
The concept of sustainable development introduces into transport department which is sustainable transport as the main development strategy for the transport department in the various countries. Sustainable transportation which includes environment, economy and society is a Multi-objective programming problem. The problem can obtains many Pareto solutions as feasible solutions through a variety of algorithm which provide decision makers to select, but decision makers directly select the ideal solution based on their own experiences and preferences in the case of many feasible solutions is difficult. Therefore, this research uses a hybrid decision making model to combine multiple objective genetic algorithm and multiple attribute decision making. Multi-objective genetic algorithms can handle complex multi-objective optimization problem to obtain Pareto solutions; multi-attribute decision making can find the preferences of different groups of decision makers and sort the Pareto solutions which helps decision maker selects a preferenced solution to carry out. In order to verify the validity of model, one example of Taoyuan city buses is used to discuss the optimization problem of sustainable city buses. Through experts questionnaires consider the views obtained from government and academic institutions, we compare the results of the hybrid decision making model and the fuzzy multi-objective programming method. The results show that city bus''s capacity is the most important evaluation criteria of experts and two model''s best compromise solutions are similar. It shows the effectiveness of the proposed model.

Dissertação de Mestrado Integrado em Engenharia Electrotécnica e de Computadores apresentada à Faculdade de Ciências e Tecnologia
Tem sido amplamente reconhecido que a adoção de medidas eficientes em termos energéticos é extremamente importante para reduzir o consumo de energia e as emissões de gases com efeito de estufa, dimininuindo também a fatura energética e aumentando a segurança energética. Além disso, o investimento em medidas eficientes em termos energéticos também implica outros benefícios relevantes que muitas vezes são negligenciados. Neste contexto, o presente trabalho procura desenvolver uma abordagem holística, considerando explicitamente múltiplos benefícios associados a várias medidas eficientes em termos energéticos. Neste âmbito, foi construído um modelo multi-objectivo, que permite obter soluções eficientes que contemplam portfolios de medidas energeticamente eficientes aplicadas ao sector residencial português. Este modelo considera cinco funções objetivo: a maximização do rácio poupança-investimento (SIR), a minimização do tempo de reembolso do carbono (CPBT), a minimização do custo da energia conservada (CCE), a minimização do risco calculado através da utilização dos pontos de vista dos diferentes peritos e a minimização da diferença para o orçamento disponível. As soluções para o modelo são então calculadas através de uma implementação ajustada baseada no Non-Dominated Sorting Genetic Algorithm. Finalmente, os resultados obtidos com esta abordagem multi-objectivo são contrastados com os calculados com uma metodologia mais próxima da tradicionalmente seguida em programas de eficiência energética. Constatou-se que numa abordagem multi-objectivo as medidas selecionadas diferem daquelas que foram obtidas com a outra metodologia, pois contemplam a análise do desempenho de ciclo de vida.
It has been broadly acknowledged that the adoption of energy efficient measures is extremely important for reducing energy consumption and greenhouse gas emissions, also lowering the energy bill, and increasing energy supply security. Besides, the investment in energy efficient measures also entails other relevant benefits that are often overlooked. In this context, the present work tries to develop a holistic approach by explicitly considering distinct multiple benefits associated with several energy efficient measures. In this framework, a multi-objective model has been built, which allows obtaining efficient solutions that contemplate portfolios of energy efficient measures applied to the Portuguese residential sector. This model considers five objective functions: the maximization of the savings to investment ratio (SIR), the minimization of the carbon payback time (CPBT), the minimization of the cost of conserved energy (CCE), the minimization of risk calculated through the use of different experts’ points of view and the minimization of the deviation from the available budget. The solutions to the model are then computed through an adjusted implementation based on the Non-Dominated Sorting Genetic Algorithm. Finally, the results obtained with this multi-objective approach are contrasted with the ones computed with a methodology closer to the one traditionally followed in energy efficiency programs. It was found that through a multi-objective approach the selected measures selected differ from the ones obtained with the other methodology, because with the former approach the life cycle performance of the measures is explicitly addressed