基于直觉模糊熵的粒子群模拟退火算法

针对智能算法在解决大规模0-1背包问题时易陷入局部最优解、收敛速度慢的问题,提出一种基于直觉模糊熵的粒子群-模拟退火算法(IFEPSO-SA)。采用交换操作和模拟退火机制对粒子群算法中的局部最优解二次优化;然后,以种群直觉模糊熵(IFE)为测度,自适应改变惯性权重,并对种群进行变异操作。测试结果表明,IFEPSO-SA在解决大规模0-1背包问题时有较好的求解质量;仿真实验结果表明,IFEPSO-SA与基于直接模糊熵的粒子群算法(IFEPSO)相比,熵值波动较小,反映出IFEPSO-SA有更好的局部搜索能力,并且IFEPSO-SA在算法收敛速度和求解质量方面都优于IFEPSO以及经典的粒子群算法和模拟退火算法。

A Hybrid Algorithm of Particle Swarm Optimization and Simulated Annealing Based on Intuitional Fuzzy Entropy

Aimed at the problem that the local optimization is easily caught in and the convergence rate is slow, this paper proposes a hybrid algorithm of particle swarm optimization and simulated annealing based on intuitional fuzzy entropy (IFEPSO-SA) in solving large scale 0-1 knapsack problems by using intelligence algorithm. Exchange operation and simulated annealing mechanism are applied to the local secondary optimization. Then, a metric based on intuitional fuzzy entropy (IFE) of the population is used to change inertia weight adaptively, and particles make the mutation based on the metric. The testing result shows that IFEPSO-SA is good in solution quality in solving large scale 0-1 knapsack problems. And the simulation experiment results show that entropy fluctuation of IFEPSO-SA is comparatively stable compared with the intuitional fuzzy entropy based particle swarm optimization (IFEPSO), reflecting a yet higher local search ability. Meanwhile, IFEPSO-SA is superior to IFEPSO and classical particle swarm optimization and simulated annealing in terms of convergence speed and solution quality.