基于改进粒子群算法的资源受限项目进度研究

资源受限的项目进度问题是经典的NP-hard问题,在研究以往求解方法的基础上,应用一种新的群智能算法——粒子群算法,对粒子群优化算法的搜索能力进行改进,结合Gbest模型与Pbest模型的优点,提出使粒子在搜索的前期有较强的全局搜索能力,尽可能多地发现可能全局最优的种子,而在搜索的后期则具有较强的局部搜索能力,用提高算法的收敛速度和精度的复合最优模型粒子群算法对RCPSP问题进行了求解,最后用文献8]中的算例进行了仿真实验,实验结果验证了此算法的可行性与优越性。

Research on resource-constrained project scheduling based on proving particle swarm algorithm

The Resource-Constrained Project-Scheduling Problem(RCPSP) is a classic NP-hard problem.Based on the analysis of former algorithms about this problem, we apply a novel swarm intelligence algorithm--Particle Swarm Optimization(PSO) algorithm,and improve it to enhance its searching ability.We combine advantages of Gbest model and Pbest model and propose that particle should be given strong global searching ability at prophase in order to find possible optimal solutions as many as possible.In contrast, particle should be given strong local searching ability at anaphase.We solve RCPSP by applying the compound optimal model PSO algorithm that improves the converging speed and accuracy of PSO algorithm.At the end of this paper,we simulate the example in reference 8],and the result proves the feasibility of this algorithm.