الفهرس 
Multi-level Multi-Objective Optimization ProblemsOnly 14 pages are availabe for public view
 |  | from | 173 |  |  |
| | | from | 173 | | |
Abstract
Optimization problems are essentially research topic in many science and engineering disciplines and there are still many open questions in this area. Multi-level optimization problems are a class of challenging optimization problems facing optimization community.
Multi-level optimization problems consist of multi interconnected hierarchical optimization problems.
Variety techniques are appeared in optimization community to solve multi-level optimization problems. Classical techniques and evolutionary techniques have appeared to solve this type of optimization problems. Classical techniques are not available to various
multi-level optimization problems especially for large scale problems. Classical techniques can be used only to specified types of multi-level optimization problems. Contrarily,
evolutionary techniques can treat different types of multi-level optimization problems.
Genetic algorithm is one of evolutionary techniques that enjoys an increasing interest in the optimization community and has a potential in tackling multi-level optimization problems.
Also, chaos is a kind of universal nonlinear phenomena in all areas of science. In the recent years, chaos theory has been applied to many aspects of the optimization sciences.
In this thesis, a modified genetic algorithm with chaotic search is proposed for solving bilevel single objective optimization problem. The integration between genetic algorithm and chaotic search procedures offers the advantages of both optimization methods. Also, other proposed algorithm is proposed to solve more difficult bi-level optimization problem, bi-level
multi-objective optimization problem. The proposed algorithm to solve bi-level multiobjective optimization problem is based on k-means cluster algorithm and genetic algorithm.
In this algorithm, we allow the decision maker to give the most his preferred solution using interactive method. The two proposed algorithms have been tested on various kinds of
benchmark problems to illustrate the proposed algorithms success.
In addition, a bilevel model is proposed for one of hardest real optimization problems, flexible job shop scheduling problem. Also, a hybrid genetic algorithm based on the bi-level
model is proposed to solve this problem. The proposed algorithm is tested on a set of standard instances taken from the literature. The computation results have validated the
effectiveness of the proposed algorithms. This thesis consists of five main chapters. These chapters can be described in the following
manner: CHAPTER 1: The most important aim of this chapter is to introduce a survey on related topics. A survey on optimization problem, multi-level optimization problem, mathematical programming techniques for multi-level optimization problems, and
evolutionary algorithms is introduced in this chapter.
CHAPTER 2: In this chapter, a new algorithm is proposed to solve bi-level single objective optimization problem. The new algorithm is a combination between modified genetic algorithm and chaotic search. The genetic algorithm has modified with a new selection
technique. Various kinds of benchmark problems have been tested to illustrate the successful result in finding optimal solution.
CHAPTER 3: In this chapter, A k-means cluster interactive algorithm based genetic algorithm for solving bi-level multi-objective optimization problems. The proposed algorithm composed of two nested artificial multi-objective algorithms. The proposed algorithm is enriched with k-means cluster scheme in two phases in the algorithm.
Various kinds of benchmark bi-level multi-objective optimization problem have been tested to illustrate the reliability of our algorithm and its ability in finding a Pareto optimal set.
CHAPTER 4: In this chapter, a bi-level model is introduced for flexible job shop scheduling problem. A hybrid genetic algorithm based on the bi-level model is proposed to solve the
flexible job shop scheduling problem. The experimental results for various kinds of benchmark problems that have been tested are discussed. The results show the proposed algorithm superiority for solving flexible job shop scheduling problem.
CHAPTER 5: This chapter describes some concluding remarks, recommendations, and some points for further research.