Dissertations / Theses on the topic '多目標規劃'

碩士
國立政治大學
應用數學研究所
92
We present an approach for the fair resource allocation problem and QoS routing in All-IP networks that offer multiple services to users. The objective of the optimization problem is to determine the amount of required bandwidth for each link and each class to maximize the sum of the users'' utility. In this work, we focus on approaches that, while allocating bandwidth, attempt to provide a proportionally fair treatment of all the competing classes. First, we will show that an achievement function can map different criteria subject to various utility onto a normalized scale. It may be interpreted as a measure of QoS (Quality of Service) on All-IP networks. Using the bandwidth allocation model, we can find a Pareto optimal allocation of bandwidth on the network under a limited available budget. This allocation can provide the so-called proportional fairness to every class, that is, this allocation can provide the similar satisfaction to each user. Next, we present a routing scheme under consideration of the delay. Such an optimal path provides the end-to-end QoS guarantees to each user. Finally, a numerical example is given to illustrate how to solve the fair resource allocation problem and how to modify the nonlinear parts.

碩士
國立臺灣大學
資訊管理研究所
91
This study focuses on the master planning of “Advanced planning and scheduling.” By considering a final product and its relationship with the global supply chain structure, the objective is to minimize the sum of production cost, processing cost, transportation cost and inventory holding cost under the constraints of limited capacities and due time requirements of orders. This planning problem, formulated as a basic model, is extended into two kinds of multiple-goal optimization problems for production planning in this study. In one extension, the extra capacities for some production nodes are allowed. With this extension, an additional objective to minimize the total amounts of extra capacities used is added into the original problem. In the other extension, the delays of orders are allowed but minimized in the added objective. In the previous study, “Linear Programming,” “Mixed Integer Linear Programming” and “Goal Programming” are some popular used to solve these kinds of problems related to supply chain management. However, with the increasing complexities of the supply chain related problems, the numbers of variables and constraints in the LP models grow rapidly. It takes a lot of computer time to solve these problems if there are feasible. Nevertheless, if the LP models result to no feasible solutions, the cause of infeasible can not be identified. Therefore, this study proposes a heuristic algorithm that is more informative and flexible then LP to solve supply chain related problems. The heuristic algorithm can indicate the status of orders and allocations of capacities and searches out feasible solutions more quickly. The heuristic algorithm first prepares the information for the supply chain by scaling the information of capacities and costs based on the unit of a final product. It then sorts all the orders to determine the sequence for planning. Finally, the orders are to be planned and scheduled one-by-one in sequence. The production plan for each order may be scheduled more than once. In each time, a minimum unit cost production tree is found at first. Following that, the available capacity of the production tree is determined. If there is no more capacity available, the supply chain network structure will be adjusted. If the cost of the schedule is larger than the one for the minimum unit cost production tree, the costs of the supply chain network will be modified and a new production tree will be found. Otherwise, the capacity of production and transportation at each node in the production tree will be allocated for this specific order. The algorithm will repeat these steps until all the orders are fulfilled. In the extension related to extra capacities, the heuristic algorithm adjusts supply chain structure by treating the nodes with extra capacities as two kinds of nodes: one physical node with regular capacity and one virtual node with extra capacity. In the extension related to order delays, the heuristic algorithm compares the schedules when order is delayed with the schedule when order is not delayed, and chooses the schedule with the lowest cost including the delay cost and the sum of the other costs.

碩士
國立成功大學
環境工程學系
87
This study applies Multiobjective Decision-Making (MODM) model on classification problems of a river basin and finds compromise strategies for trading off among conflicting objectives that cannot be solved by single objective programming techniques. Four major categories considered in the objective function are water quality in a river basin, allowable pollution loading, treatment cost of wastewater and the benefits of water use. The objective of water quality in a river basin includes the minimum concentrations of BOD5 and total phosphorous, and the objective of allowable pollution loading contains the maximum quantity of BOD5 and total phosphorous, too . In this study, Tweng-Wen River basin was selected as a case study in which six objectives and over ten constraints are included accordingly. The MODM model can determine the best compromise classification of each reach in Tweng-Wen River using the multiobjective simplex method.

碩士
國立勤益科技大學
工業工程與管理系
104
Assembly line balancing problem is often encountered in a manufacturing system. In order to maintain a certain level of production rate, several objectives need to be considered at the same time. However, these objectives are not necessarily certain in real environments, and thus, a fuzzy multiple objective assembly line balancing problem （FMOALBP） is present. In traditional mathematical models, a single objective is considered, and the establishment of a model is to promote the quality of the objective. Therefore, such a model cannot solve a complicated problem. In this paper, a fuzzy multi-objective piecewise linear model for the assembly line balancing problem is applied, and there are four objectives: minimizing cycle time, minimizing the number of workstations, minimizing workload smoothness, and maximizing product profits. This multiple-objective model is converted into a single-objective model, and the piecewise linear membership function is used to calculate the satisfaction degree. A case study is presented to examine the model. And when case size is complicated, exhaustive enumeration become impractical to application. Therefore, heuristic may be necessary to solve the problem. In this paper, genetic artificial immune systems algorithm is constructed for solving large-scale scheduling problem and utilizes it to improve efficiency and quality of solution. In a real practice environment, uncertain condition and fuzzy objective are often present. The purpose of this paper is to help decision makers consider the overall problem objectively and then make satisfactory decisions.

碩士
國立勤益科技大學
工業工程與管理系
104
The improved gross domestic product and living standards have caused arising demand in housing quality: the function of housing has progressed from a place of shelter to the modern demands of housing quality and public facilities. However, consumer demand for ideal housing has been reprioritized to value not merely quality, which was formerly the primary concern, but also additional factors such as location, neighborhood, and local transportation options. Construction groups should learn and know how to fulfill consumer demand because unlike fast-moving consumer goods, housing decision making does not typically offer immediate replacement. Housing purchases should accommodate personal demand, as reflected in housing prices. Therefore, the objective of this research was to provide an instant housing demand evaluation method based on consumer demand. Generally, housing decision making involves a minimum of two decision-makers. Group decision-making uses an integrated means to reach an agreement on the basis of the mutual consensus of all participating decision-makers. Such an agreement should be a common goal for each consumer and construction group. When confronting a critical decision, decision-makers frequently lack a sense of certainty and clarity when communicating their demands. In this study, fuzzy theory was used as a primary demand communication tool and multi-choice goal programming was employed to assist decision-makers to conduct an integrated data analysis. This study used Borland C++ Builder 6.0 as a development platform to integrate the applied methodologies to digitalize and expedite decision-making. The findings revealed that each decision-maker required an average of 5 minutes to list his or her ideal housing criteria and demands, thus facilitating effective and satisfactory housing selection.

碩士
國立臺灣大學
環境工程學研究所
89
Environmental Engineer paid much attention to air pollution and water pollution in early periods and only until recently did they focus on groundwater pollution. Goundwater pollution is a complex environmental problem with high uncertainty. Under the situation without enough information, it is important for decision makers to obtain guidance on the decision problems. The main objective of this study is to develop a method of making decisions under uncertainty in a groundwater-contamination context. Since traditional deterministic programming can’t deal with uncertainty, this study uses stochastic programming and fuzzy programming to address the influence of uncertainty on groundwater remediation. Furthermore, multiobjective programming considering cost and health risk is incorporated with uncertainty analysis in this study to select optimal alternative with minimal health risk and remediation cost. If there is enough information on probability density functions, stochastic programming should be the preferred approach; If no enough information is available, fuzzy programming should be used instead. The programming results of weighting method at Wcost：Wrisk=0.5：0.5 under uncertainty change from air sparging(b) to bioremediation(a) and then to natural attenuation with adding confidence levels. The results of compromise programming under uncertainty are similar, except that for compromise programming combined with Zimmermann method there is no feasible solution. The result derived from other combinations of fuzzy programming is air sparging(b), with membership degree between 0.4608 to 0.6433.