Improved Analysis And Applied Research Of Artificial Glowworm Swarm Optimization Algorithm

In many scientific areas and engineering computation areas, most of problems that people encounter can be attributed to objective optimization problem. With the rapid growth and wide application of electronic computer, optimization technique gets the rapid development and has long been a focus for researchers all the time. In recent years, with the rapid development of computational intelligence theory and technology, people has proposed a variety of social bionic evolution algorithms, including genetic algorithm (GA), ant colony algorithm (ACA), particle swarm optimization (PSO), simulated annealing (SA) and so on. For solving many complex optimization problems, these algorithms have demonstrated their excellent performance and great development potential. So the intelligent algorithm research is a subject with important theoretical significance and practical application value.Artificial glowworm swarm optimization (GSO) algorithm that simulated intelligent behavior of glowworm swarm in nature is a new intelligence optimization algorithm. GSO is simple and is also implemented easily. In addition, it has strong robustness and easily integrates with other algorithms. So far, GSO has been successfully applied to detection of multiple source locations, multimodal function optimization, network robot system and so on. However, it also has some shortcomings for searching the global optimal solution, such as the slow convergence speed, easily falling into the local optimum value, the low computational accuracy and success rate of convergence.In view of the defects of GSO, the main researches are as follows in this paper:(1) According to the chaotic motion with randomness, ergodicity and intrinsic regularity, chaos method as a local search operator is embedded into GSO and a hybrid artificial glowworm swarm optimization algorithm is proposed. That is to say, in the evolutionary process, firstly, GSO implements the global search. Then, chaos method implements the local search within the given number of steps for the glowworms whose current fitness values are better than the average fitness value, which leads the swarm to the direction of the optimal solution. So the hybrid algorithm ensures global convergence and local ergodicity, and it is easier to escape the local optimum value.(2) In view of the powerful local optimization ability of Powell method, this paper embeds it into GSO as a local search operator and proposes an artificial glowworm swarm optimization algorithm based on Powell local optimization method. Experimental results show that the proposed algorithm is far superior to GSO in convergence efficiency, computational precision and stability.(3) Predatory behavior of artificial fish swarm algorithm (AFSA) is embedded into GSO, that is to say, the glowworms whose neighbor sets are empty are carried out predatory behavior in their dynamic decision domains. Then the improved GSO is combined with differential evolution (DE) on the basis of an optimal information sharing and two-population co-evolution mechanism, a hybrid optimization algorithm is proposed. The experimental results based on some typical functions and a minimum thrust ball guide engineering optimization design show that the hybrid algorithm has better convergence efficiency, higher computational precision and better the performance.(4) Aiming at constrained optimization problems, a hybrid artificial glowworm swarm optimization algorithm is proposed. The proposed algorithm takes GSO and DE as the basic framework, updates the optimum position of the population using updating strategy on the basis of the feasibility rules in the search process, and adopts the local search strategy based on SA for the optimal position of each generation. Finally, the presented algorithm is tested on some well-known benchmark functions and five engineering optimization functions. Comparisons show that the proposed algorithm has faster convergence efficiency, higher computational precision, better robustness and is more effective for solving constrained optimization problem.