Differential Evolution Algorithm And Application Research In Route Planning For Unmanned Air Vehicles

As optimization problems exist widely in all domains of scientific research and engineering application, research on optimization methods is of great theoretical significance and practical value. Therefore, effective optimization methods have become one of the main objectives for scientific researchers.Differential evolution algorithm (DE) is an evolutionary computation technique developed by Storn and Price in 1995. It has some features such as easy to use, simple in concept, fast convergence speed and need little information about problems. It is proved to be very suitable to solve some complex optimization problems and has successfully been applied in the area of function optimization, neural network training and fuzzy control systems, etc. But, DE algorithm is easy to fall into a locally optimized point, because of lacking mechanism of using Global information.The Particle Swarm Optimization (PSO) method was originally designed by Kennedy and Eberhart in 1995 and has been applied successfully in various optimization problems. PSO algorithm possesses a strong mechanism for global search because of using Global information. Sometime, PSO algorithm is easy to be in premature convergence as a result of loss of population diversity. The differential evolution algorithm has many advantages, such as maintaining the population diversity, as well as good local search ability. But it lacks the guidance of global information, so easily lead to a waste of computing resources and be in premature convergence. So they can make the integration.This article leads the multi-agent system technology into the DE algorithm and the'local version'DE algorithm is proposed. In order to overcome the defect of DE and PSO algorithm in solving a global optimization problem, the paper proposes a novel global optimization algorithm, Co-evolutionary Differential Evolution (CDE). CDE algorithm based on a two-population evolutionary strategy, in which the individuals of one population are evolved by PSO and the individuals of the other population are evolved by DE. The individuals both in PSO and DE are co-evolved during the algorithm execution by employing information sharing mechanism, to avoid their own population into a local optimum. The method can guarantee the capacity that the role of the two populations with each other to coordinate the development of balance capacity between development and exploration.The CDE algorithm has been applied to route planning for unmanned aerial vehicle. The simulation result has indicated that CDE was able to undertake global search with a fast convergence rate and a feature of robust computation. It was proved to be validity, fast convergence and computation efficiency during the route planning.