多目标资源优化分配问题的Memetic算法

Hybrid effective Memetic algorithm for multi-objective resource allocation problem

针对传统算法求解多目标资源优化分配问题收敛慢、Pareto解不能有效分布在Pareto前沿面的问题,提出一种新的Memetic算法.在遗传算法的交叉算子中引入模拟退火算法,加强了遗传算法的局部搜索能力,加快了收敛速度.为了使Pareto最优解均匀分布在Pareto前沿面,在染色体编码中引入禁忌表,增加了种群的多样性,避免了传统遗传算法后期Pareto解集过于集中的缺点.通过与已有的遗传算法、蚁群算法、粒子群算法进行比较,仿真实验表明了所提出算法的有效性,并分析了禁忌表长度和模拟退火参数对算法收敛性的影响.

For the multi-objective resource allocation problem（MORAP）, the traditional algorithms execute slowly and non-dominated solutions can’t uniformly distribute in the Pareto front. Therefore, a new memetic algorithm（HTGSA） is proposed. The simulated annealing algorithm is introduced to GA to strengthen the local search and convergence speed. To make the non-dominated solutions uniformly distribute in the Pareto front, a tabu constraint strategy is introduced to the survival selection process of genetic algorithm. The tabu constraint strategy can strengthen the diversity of the solutions and prevent non-dominated solutions over concentration in the Pareto front. Finally, the proposed algorithm is tested with the numerical simulation experiment comparing the results with ACO, GA and PSO. Experimental results on MORAP show that this hybridization can significantly accelerate the convergence speed and reduce the computation time.