多模式资源受限项目调度问题的优化方法

An Optimal Approach for Solving Multi-mode Resource-constrained Project Scheduling Problem

为求解多模式资源受限项目调度问题,提出了一种结合粒子群优化算法(PSO)和基因表达式编程(GEP)的混合优化算法。其中,PSO用来提供活动执行模式组合,GEP用来构造在给定活动执行模式下的调度规则。调度规则由项目状态和活动属性构成,与其他优化方法相比,这是一种新的编码方式与求解方法,也更符合实际应用。对于粒子所表达的不可行活动模式的组合,设计了粒子调整算法,以满足项目调度中可更新资源和不可更新资源总数的约束。最后给出了混合优化算法求解步骤,并采用该算法对项目实例进行了计算与分析,验证了算法的有效性。

In order to solve multi-mode resource-constrained project scheduling problem （MMRCPSP）, a hybrid optimization method integrated particle swarm optimization （PSO） with gene expression programming （GEP） is proposed, where PSO is used to search the mode combination for all activities and GEP is applied to construct the effective scheduling rule to determine the priority of activities with given mode combination. The scheduling rule is the algebraic combination of project status and attributes of activity, which is a new representation form of solution compared with other optimal algorithms for MMRCPSP and is more suitable for the real- world application. The procedure to check and adjust the infeasible particle-represented mode combination is designed to meet the limi- tations on the total number of renewable and non-renewable resources. At last, the framework of the proposed method for MMRCPSP is introduced and experimental analysis is presented to investigate the performance of the method.