基于并行策略的改进混合粒子群算法及其应用

Improved Hybrid Particle Swarm Optimization Algorithm Based on Parallel Strategy and Its Application

针对粒子群算法在解决高维度复杂优化易陷入局部最优的问题,构建差分进化算法(DE)、人工蜂群算法(ABC)与粒子群算法(PSO)并行运算的种群更新模型,提出基于并行策略的改进混合粒子群算法(DA_PSO)。以并行策略为基础,不改变种群规模,独立运行3种算法,每隔n次比较3种算法,获得当前最优点,并用其替换粒子群算法的种群最优点,利用PSO算法个体向种群最优靠近的特点,充分吸收DE算法、ABC算法的优点,使被替换后的PSO算法跳出局部最优,提升优化结果的质量。采用五种类型测试函数分别对ABC、DE、PSO和DA_PSO进行对比验证,结果表明:较其他算法而言,DA_PSO算法精度高,稳定性好,适应性强。同时为验证所提方法的科学性与实用性,将其应用在10t~32t/31.5m系列化的桥式起重机主梁金属结构轻量化设计中。

In order to solve the problem that particle swarm optimization is easy to get into local optimal in high dimensional complex optimizations,we construct a population updating model of parallel operation of difference evolutionary algorithm(DE),artificial bee colony algorithm(ABC)and particle swarm optimization(PSO),and propose an improved hybrid particle swarm optimization(DA_PSO)algorithm based on parallel strategy.Based on the parallel strategy,we do not change the population scale,run 3 kinds of algorithms independently,compare 3 kinds of algorithms every n times,obtain the most advantage of the current,and replace the most advantage of the population with the PSO algorithm,utilize the PSO algorithm’s characteristics of the optimal proximity of the individual to the population,and fully absorb the advantages of the DE algorithm and ABC algorithm.The proposed PSO algorithm jumps out the local optimum and improves the quality of the optimized result.Using five types of test functions to compare and verify ABC,DE,PSO and DA_PSO,the results show that:compared with other algorithms,the DA_PSO algorithm has high precision,good stability and strong adaptability.In order to verify the scientific and practicability of the proposed method,it is applied in the lightweight design of the metal structure of the main girder of the 10t~32t/31.5m series Bridge Crane.