Improved Differential Evolution Algorithm Based On Strategy Of Control Parameter Co-evolution And Its Application

In the field of control science and engineering, optimization problem have always been a research hotspot.With the development of science and the deep understanding object, the characteristic of optimization object is high-nonlinear, high-dimension and complex. Thus, it is difficult to solve these optimization problems or obtain the global optimum solution by traditional optimization technology. With the development of intelligent optimization algorithms, which can obtain global optimum solution with probability, are widly used for complex opitimization problems. Differential evolution algorithm is a kind of intelligent optimization algorithms, and has many merits such as fast convergence rate, good robust property and high probability of global convergence. The performance of DE is mainly dependent on the value of control parameters; meanwhile, self-adaptive adjustment of control paremeters with population evolution to provide the suitable control parameters for current population is also important for its performance. Therefore, differential evolution algorithm with control parameter co-evolution, which implements self-adaptive control paremeters by co-evolving strategy to improve the convergence rate and the probability of global optimal seach, is researched. The algorithms are applied to solve parameters optimization, constrained optimization and dynamic optimization. The results show that the proposed algorithms are effective and superior.(1) A novel differential evolution algorithm with control parameter co-evolution (DE-CPCE) is proposed. In DE-CPCE, control parameters are designed as the symbiotic individual of original individual, and each original individual has its own symbiotic individual. Differential evolution operator is applied to search the global optimization solution of problem; meanwhile, it is also employed to co-evolve the population consisting of symbiotic individuals according to the evolution efficiencies of original individuals. DE-CPCE is applied to test the benchmark functions and estimate the kinetic model parameter of so2 oxidation. The results show that performance of DE-CPCE is better than SPDE algorithm and SDE algorithm and so on. (2) Considering that it is difficult to set suitable penalty factors for the penalty function method, a novel differential evolution algorithm with co-evolution of control parameters and penalty factors, named as CoE-DE, is proposed. In CoE-DE, differential evolution operator is applied for evolving the original individuals. To improve the performance of CoE-DE and the handling constraints capacity, Alopex algorithm is used to co-evolve the symbiotic individuals, which consist of two DE control parameters and the penalty factors. SADE is applied to solve the benchmark functions and 5 constrained engineering problems, the results show that the performance of CoE-DE is better than SR algorithm and similar to a SIMPILE.(3) In order to solve dynamic optimization problem of batch reactor, a self-adaptive differential evolution algorithm (SADE) is introduced. In SADE, each original individual has its own control parameters. Differential evolution operator is employed to search the optimization results of problems, and weight value is applied for evaluating the corresponding control parameters. Meanwhile, the weighted average of the control parameters, which is calculated based on the weight value and the corresponding control parameters, is used for guiding the evolution direction of control parameters. SADE is applied to solve the benchmark functions and two typical dynamic optimization problems of batch reactor, the results show that the performance of SADE is better than JADE, IACA algorithm and so on.(4) To solve dynamic optimization problem of chemical process, a hybrid differential evolution algorithm, which is integrated with Alopex (Alopex-DE), is proposed. In Alopex-DE, each original individual has its own symbiotic individual consisting of control parameters. Differential evolution operator is applied for the original individuals to search the global optimization solution. Alopex algorithm is used to co-evolve the symbiotic individuals during the original individual evolution and enhance the fitness of the original individuals. Alopex-DE is applied to solve the benchmark functions and dynamic optimization problem of chemical process. The results show that the performance of Alopex-DE is better than SPDE, SDE, FEP algorithm and so on.