Research Of Differential Evolution Algorithm Applications On Multi-Objective Optimization

In scientific and economic fields, many problems can be attributed to multi-objective optimization problem. In the multi-objective problem, the objectives often do not existence independent but rather conflicting. As a result, a single vector that makes all the targets optimum does not exist. This makes it difficult to optimize the multi-objective problem. Differential evolution algorithm is a real-coded optimization algorithm. It is simple in principle, controlled by few parameters, easy to understand and realize. It can deal with complex function optimization, global optimization and multi-objective optimization problems effectively. In recent years, the differential evolution algorithm is applied to multi-objective optimization problem, and it has become a new research direction.In this paper, the research status, research methods, progress of research, and hot applications of differential evolution algorithm and multi-objective optimization are systematically summarized. The difficulties that differential evolution algorithm field encountered currently were analyzed. All this has provided guidance to improve the algorithm. Meanwhile, this paper reviews some representative ideas of differential evolution algorithm applied to the multi-objective problem.Multi-objective diet decision-making in the optimization process is prone to the phenomenon of optimum populations is very few, therefore the definition of similar distance is raised, and the similar distance being added to the trade-offs differential evolution algorithm. Then the improved differential evolution algorithm is applied to multi-objective optimization decision-making in the diet. At the same time, population diversity is increased by the improved algorithm. Numerical simulation experiments have been carried out in matlab. The experimental result shows: the use of the improved differential evolution algorithm obtained Pareto solution is more accurate than the use of genetic algorithms.By studying of the multi-objective distribution vehicle routing optimization with a time window, this paper has improved the mutation operator of the differential evolution algorithm, and applied this algorithm in more complex multi-objective optimization problems successfully. The numerical simulation result shows: the differential evolution algorithm with improved mutation operator can get more effective non-dominated solution.