Research On Multi-objective Different Evolution Algorithm Based On Jumping Genes

In the process of scientific research and engineering design, many specific problems can be summarized as parameter optimization problem, in the reality, these optimization problems often have multiple goals, the goals contradict and restrict each other, The improvement of the performance of a optimization goal often leads to at least one other objective performance degradation, so multiple target is very difficult to achieve optimal at the same time. Therefore, multi-objective optimization algorithm research has been a hot spot of today's science and engineering design research direction. Evolutionary algorithm is a generic terms of heuristic search and optimization algorithm, which is inspired by biological nature and system, using evolutionary algorithm to solve multi-objective optimization problem, It has been widely used. As an important part of the evolutionary algorithm, differential evolution algorithm is easy to understand, and have simple structure,less adjustable parameter, the robust intelligent optimization method.DEMO(Differential evolution algorithm for multi- objective optimization) and MODEA(Multi-objective Differential evolution algorithm), etc, are known as famous algorithms In Multi-objective Differential algorithm. However, DEMO and MODEA are both very easy to fall into local optimum when solve complex problems, due to DEMO uses a strategy of rapid father exchange and MODEA uses a greedy mutation strategy.In order to handle the problem that the existing multi-objective differential evolution algorithms are easy to get the local optimum result, the jumping genes operation is introduced and a novel jumping genes based multi-objective differential evolution algorithm(JGMODE) is presented. Different from the existing algorithms such as DEMO and MODEA, JGMODE algorithm performs jumping genes operation after the classical crossover operation, in order to improve the population diversity and thus enhance the exploration ability of the algorithm. Numerical experimental results indicates that the proposed algorithm is capable of dealing with the local optimum problem and exhibits significantly better performance than the existing algorithms on difficult test problems included in the ZDT and DTLZ test suits.