机电安装工程成本–进度协同优化研究

Integrated Optimization of Cost and Schedule in Mechanical and Electrical Installation Project

摘要: 为实现工程项目管理中工期与成本的协调优化，本文提出一种融合Q-learning强化学习与粒子群优化(PSO)算法的多目标优化方法。通过构建带权有向图模型，结合改进的Q-learning算法动态识别关键路径，并引入日间接成本率自动计算机制，构建工期与成本耦合的综合效益函数。并利用PSO算法对工序工期进行全局优化，有效提升项目综合效益。以实际工程数据为例，分析了优化前后关键路径、工期及成本变化，并开展工期权重与惩罚参数的敏感性分析，揭示参数对优化结果的影响。结果表明，模型具备良好稳定性和适应性，为项目管理提供了智能化优化支持。

Abstract: To achieve coordinated optimization of time and cost in engineering project management, this paper proposes a multi-objective optimization approach integrating Q-learning reinforcement learning and Particle Swarm Optimization (PSO). A weighted directed graph model is constructed, in which an improved Q-learning algorithm is employed to dynamically identify the critical path. An automatic estimation mechanism for the daily indirect cost rate is introduced to formulate a coupled time-cost comprehensive benefit function. The PSO algorithm is then used to globally optimize activity durations, effectively enhancing overall project performance. Using real engineering data as a case study, the changes in critical path, project duration, and cost before and after optimization are analyzed. A sensitivity analysis is also conducted on the time weight and penalty parameters to reveal their impact on the optimization results. The results demonstrate that the proposed model exhibits strong stability and adaptability, offering intelligent decision-making support for project management.