【摘要】：針對服務(wù)質(zhì)量(quality of service,QoS)全局最優(yōu)Web服務(wù)選擇問題,提出了一種雙種群協(xié)同進(jìn)化QoS全局最優(yōu)Web服務(wù)選擇算法。算法在多目標(biāo)離散粒子群算法基礎(chǔ)上設(shè)計(jì)一種雙種群協(xié)同進(jìn)化框架以同步進(jìn)行非支配排序和精英粒子保留,并定義了一種新的離散粒子位置更新算子。同時(shí)為保證粒子的多樣性和算法的全局收斂能力,算法采用基于距離的粒子多樣性度量算子、基于適應(yīng)值排序的粒子選擇算法和基于輪盤賭的全局最優(yōu)解選擇策略。仿真實(shí)驗(yàn)結(jié)果表明該算法能同時(shí)優(yōu)化多個(gè)目標(biāo),并得到一組滿足約束的Pareto最優(yōu)解,且具有較好的性能和魯棒性,解集的質(zhì)量和分布也優(yōu)于非支配排序遺傳(nondominated sorting genetic algorithm,NSGA)算法的改進(jìn)算法NSGA-Ⅱ,能有效解決QoS全局最優(yōu)的Web服務(wù)選擇問題。
[Abstract]:Aiming at the globally optimal (quality of service selection problem, a two-population coevolutionary QoS global optimal Web service selection algorithm is proposed. Based on the multi-objective discrete particle swarm optimization algorithm, a two-population coevolutionary framework is designed to synchronize non-dominated ordering and elite particle retention, and a new discrete particle location update operator is defined. In order to ensure the diversity of particles and the ability of global convergence, the algorithm adopts a distance based particle diversity metric operator, a particle selection algorithm based on fitness ranking and a global optimal solution selection strategy based on roulette. Simulation results show that the algorithm can optimize multiple targets at the same time, and obtain a set of Pareto optimal solutions which satisfy the constraints, and have good performance and robustness. The quality and distribution of the solution set is also superior to the improved algorithm NSGA- 鈪，

本文編號：2185990