The Research Of Global Intelligent Optimization Algorithm

With the continuous development of economy and computer technology, aword-”optimization” has penetrated into our daily lives and been fully demonstrated inincreasingly complex engineering applications. Because of today’s engineering applicationshaving characteristics such as large-scale, multi-task, strong mobility, etc, traditionaloptimization methods tied to their own limitations are difficult to solve these problems. Withthe continuous development and application of artificial intelligence, intelligence optimizationalgorithms have strong advantage in solving complex engineering optimization problems,having become hot topic in the field of artificial intelligence and playing an increasinglyimportant role.Particle swarm optimization(PSO) algorithm is a typically global swarm intelligenceoptimization algorithm, and it has caused serious attention and research of many academicsand experts at home and abroad because of its simple concept, less parameters, easyimplementation, strong optimization ability, etc. However, as a kind of heuristic swarmintelligence optimization algorithm, the theoretical basis of PSO is still far from mature, lackof mature demonstration and research, leading many spaces to upgrade and improve in theengineering application. The paper gives a comprehensive study on PSO from the aspect ofalgorithm mechanism, modification and its application.This paper based on theory of PSO, analyses the track of particle and disadvantage ofgetting struck at local optima easily and appearance of premature convergence, the regionalshock search embedded particle swarm optimization(RSPSO) algorithm is proposed. RSPSOsets attractor of each particle in the populations as center, and finds the best solution infeasible solution space by shock search around the attractor. The algorithm takes region shocksearch strategy to enhance the diversity of the population and prevent premature convergence.Meanwhile, The algorithm executes around attractors of population, without the need for PSOto update velocity of particle, making the algorithm less parameters to adjust and easier toimplement. Verified by typical test functions, the RSPSO algorithm shows better solutionaccuracy during the process of optimization.Resource-constrained project scheduling problem(RCPSP) widely exsits in variousindustries in real life, having been proved to be NP-hard. Because of the difficulty of solving,it attracts many scholars at home and abroad to analyse. For solving RCPSP, the CooperativeShock search Particle Swarm Optimization with Chaos (CSCPSO) which absorbs the idea andfeature of RSPSO is proposed in the paper. On basis of particle attractor, a bidirectionalcooperation shock search mechanism is established in the algorithm to enhance the searchaccuracy and diversity of population. Particles converge to particle attractor, meanwhile theyadjust dimensions whose adjacent relationship are inconsistent with attractor’s by shocksearch in the mechanism. Combined with topological sorting based on particles and serialschedule generation scheme, the gotten scheduling scheme can meet the project scheduleconstraints of resource and precedence relations. The tests on specific examples show that theproposed algorithm can get higher accuracy and better stability for RCPSP and will be more usefull for practical application.