Comparing Heuristics for Solving Linear Bilevel Problems

Florian Grodeke

ISBN 10: 3668266336 / ISBN 13: 9783668266339

Published by GRIN Verlag, 2016

About this Item

Druck auf Anfrage Neuware - Printed after ordering - Bachelor Thesis from the year 2013 in the subject Business economics - Operations Research, grade: 1,0, Technical University of Munich (Lehrstuhl für Betriebswirtschaftslehre - Logistik und Supply Chain Management), language: English, abstract: In this thesis, metaheuristic solution approaches for the decentralized capacitated facility location problem are deduced, implemented and assessed. As the mentioned problem is a combinatorial bilevel problem it is not easily solvable by means of linear programming. Therefore, different metaheuristic techniques, namely Tabu Search and Simulated Annealing, are used. Each of them is split into several variants. Tabu Search has turned out to deliver better results than Simulated Annealing and converges also from bad initial solutions towards the global optimum. However, the performance of the algorithms is highly dependent on the concrete parameter settings. Thus, the gap between the optimum found and the global optimum can vary from 0 to a multiple of the global optimum depending on the used variant. The thesis on hand describes the comparison of different metaheuristics in order to generate solutions for the bilevel decentralized capacitated facility location problem. This optimization problem consists of choosing a set of facilities to be opened out of a superset of potential facilities. The objective is to minimize cost, while taking certain constraints into account. This usual facility location model, which can relatively easily be solved as a linear optimization problem, will be expanded. There is an additional submodel that is intended to develop an optimal production plan for the facilities chosen within the superordinate model. On the sublevel, also cost has to be minimized and as one major constraint the entire external demand has to be satisfied. Such a bilevel problem containing two hierarchically arranged objective functions cannot be solved using the usual methods of linear programming any more.

Seller Inventory # 9783668266339

Bibliographic Details

• Title: Comparing Heuristics for Solving Linear Bilevel Problems
• Publisher: GRIN Verlag
• Publication Date: 2016
• Language: English
• ISBN 10: 3668266336
• ISBN 13: 9783668266339
• Binding: Taschenbuch
• Condition: Neu
• Dimensions: 210W x 148H x 6D millimeters 210 millimeters width by 148 millimeters height by 6 millimeters depth
• Weight: 112 grams

About this title

Synopsis

Bachelor Thesis from the year 2013 in the subject Business economics - Operations Research, grade: 1,0, Technical University of Munich (Lehrstuhl für Betriebswirtschaftslehre - Logistik und Supply Chain Management), language: English, abstract: In this thesis, metaheuristic solution approaches for the decentralized capacitated facility location problem are deduced, implemented and assessed. As the mentioned problem is a combinatorial bilevel problem it is not easily solvable by means of linear programming. Therefore, different metaheuristic techniques, namely Tabu Search and Simulated Annealing, are used. Each of them is split into several variants. Tabu Search has turned out to deliver better results than Simulated Annealing and converges also from bad initial solutions towards the global optimum. However, the performance of the algorithms is highly dependent on the concrete parameter settings. Thus, the gap between the optimum found and the global optimum can vary from 0 to a multiple of the global optimum depending on the used variant. The thesis on hand describes the comparison of different metaheuristics in order to generate solutions for the bilevel decentralized capacitated facility location problem. This optimization problem consists of choosing a set of facilities to be opened out of a superset of potential facilities. The objective is to minimize cost, while taking certain constraints into account. This usual facility location model, which can relatively easily be solved as a linear optimization problem, will be expanded. There is an additional submodel that is intended to develop an optimal production plan for the facilities chosen within the superordinate model. On the sublevel, also cost has to be minimized and as one major constraint the entire external demand has to be satisfied. Such a bilevel problem containing two hierarchically arranged objective functions cannot be solved using the usual methods of linear programming any more.