Hybrid Artificial Glowworm Swarm Optimization Algorithm And Its Applications

Artificial glowworm swarm optimization algorithm (GSO) is a novel swarmintelligence optimization algorithm, which is inspired by glowworm’s glowingbehavior of foraging or courtship in nature. GSO has been successfully employedto solve function optimization and other engineering problems, such as multiplesignal sources localization. However, with the deepening of the research, theresearchers found that GSO own many defects and shortcomings，including slowconvergence speed, low computational accuracy and usually lost in local optimum.In view of these shortcomings, this thesis has conducted in-depth analysis, anddevised three mixed GSO algorithms which are applied to solve functionoptimization or some practical engineering problems. The experimental resultsshow that the proposed hybrid algorithms are feasible and effective.This thesis mainly achieved the following research results:(1) Based on the characteristics of GSO and parallel hybrid mutation ideas, thestrategy of parallel hybrid mutation is introduced into GSO and a hybrid algorithm(GSO based on parallel hybrid mutation, PHMGSO) is proposed. In order to testthe performance of PHMGSO, ten classical functions were tested. Experimentalresults show that PHMGSO own higher calculation accuracy and fasterconvergence speed than GSO.(2) GSO and Particle swarm optimization (PSO) are integrated together andan improved GSO based on PSO (PGSO) was proposed. In PGSO, the update ofindividual location both includes the local information and global optimalinformation. In this way, the cost of individual exchange information can bereduced. In order to verify the performance of the PGSO, PGSO is applied to solveuninhabited combat air vehicle (UCAV) path planning problem. Simulation resultsshow that PGSO is effective to solve UCAV path planning problem.(3) Based on the quantum computation and the characteristics of GSO, thisthesis presented QGSO algorithm. QGSO contains the idea of quantum coding,interference and variation. Furthermore, QGSO is applied to solve theone-dimensional packing problem. The simulation experiment results show that the QGSO algorithm is effective on one-dimensional packing problem.