基于粒子群算法的资源受限项目扩展调度方法

An Extended Scheduling Approach Based on Swarm Optimization Algorithm to Resource-constrained Project Scheduling Problem

针对资源受限的项目调度问题,提出了一种离散粒子群算法与扩展调度机制相结合的优化方法.离散粒子群算法中每个粒子的位置代表一组项目任务的优先权,迭代中通过交叉策略和局部搜索策略来更新粒子的位置,这既保持了粒子位置的离散性,又增加了粒子的多样性,避免早熟收敛.每个粒子的位置通过扩展串行调度机制转换成可行的调度方案.实算表明,扩展调度机制的引入显著地加速了收敛的进程,提高了解的精度.这种基于粒子群算法的扩展调度优化方法是求解资源受限项目调度问题的有效方法.

To solve the resource-constrained project scheduling problem（RCPSP）,an optimization algorithm combined discrete particle swarm optimization（DPSO） with extended serial scheduling scheme is presented.In DPSO the location of each particle represents priorities of activities.Cross strategy and local search technology are adopted when updating the particle positions,which can ensure discreteness,increase the diversity of particles and avoid premature during iteration.The location of each particle is transformed to a feasible schedule by the extended serial scheduling scheme.The actual calculation shows that extended scheduling scheme can accelerate the convergence process obviously and improve the precision.It can be concluded that this algorithm is valid for RCPSP.