Research Of Differential Evolutionary Algorithms Application On Combination Optimization

It is common to face a lot of combinatorial optimization problems in the real world.Usually,these problems are often constrained. So it is hard to solve these problems efficiently by using traditional methods.As a new natural method and robust evolutionary algorithms based on population,Differential Evolution Algorithms are promising to solve the constrained combinatorial optimization problems.But the precision and efficiency of the algorithm are still faced with the bottleneck.Therefore,discussing and designing the effective algorithms for the specific optimization problem has received more and more attentions in the world.And it is gradually becoming the research focus in the field of evolutionary computation.This study centering on the classcial knapsack problem and congruent circle problem, which are two kinds of typical combinatorial optimization problem,are deeply summarized and get some research results as follows:(1) For the combinatorial optimization problem of discrete space,a Permutation-Based Differential Evolutionary algorithm is proposed by using the Permutation strategy.Main works are summarized as follows:1)a new algorithm of Permutation-Based discrete Differential Evolution is proposed, which absorbs the Permutation strategy and Differential Evolutionary algorithm.And the algorithm is applied to the single-object knapsack problem solving.2)a novel multi-objective differential Evolutionay is proposed by using elite retain mechanism NSGA-Ⅱand is implement within the framwork of Differential Evolutionary.3)a novel algorithm of Permutation-based Multi-objective discrete Differential Evolution is proposed and it has been applied to solve the Multi-objective knapsack problem.Experimental results and the analysis on the knapsack problem demonstate the effectiveness of the algorithm. Especially it is superior to classical NSGA-II algorithm in large-scale problem.(2)For the combinatorial optimization problem of continuous space with large-scale constraints, a appropriate differential evolution algorithm is designed and a lot of experiments have been done to prove the effectiveness of the algorithm, In this dissertation,research focusing on the Congruent Circle Packing Problem is discussed. And a novel differential evolution algorithm has been designed for it. In the novel algorithm the Congruent Circle Packing Problem is converted into multi-objective problem.Experimental results can prove that this algorithm is effective.In this dissertation,with the studies in combinatorial optimization problem. A novel Permutation-Based discrete Differential Evolution algorithm are designed for discrete optimiztion space, and a novel differential evolution for Congruent Packing problem is designed.The works in this dissertation are very important to the research of the differential evolutionary optimization and the optimization applications of DE in the real world.