用于施工进度优化的VNS-CRO混合算法

Hybrid algorithm of VNS-CRO for construction process optimization

针对现有研究中忽略施工任务分包资源系统异构的问题,结合建筑信息模型技术与异构环境下任务调度理论,提出变邻域搜索-化学反应混合优化算法,实现了对异构资源环境下道路施工进度计划任务调度与分包的优化.该算法在有向无环图模型与异构资源系统模型的基础上,融合了约束关键路径重排序策略,并在初始化与优化构造超级分子阶段,设计应用了一种平衡邻域结构的变邻域搜索方法,强化了算法的整体优化能力.模拟试验使用上海市陈翔路工程的实际工程数据,试验结果验证了所提出算法优化的有效性与鲁棒性及其相对其他算法的优势.

To solve the absence of heterogeneous resource systems for construction task sublet and schedule in current research works, a hybird algorithm of variable neighbourhood search-chemical reaction optimization （VNS-CRO） was proposed. The building information modeling （BIM） technology was combined with the theory of task scheduling on heterogeneous systems to construct task schedule arid sublet on heterogeneous resource systems. Based on directed acyclic graph （DAG） model, heterogeneous resource system model and constrained critical path strategy, the variable neighbourhood search method with balanced neighbourhood structure was utilized in initiation and optimizing super molecule for improving optimization capability. The data of Chengxiang Road construction project in Shanghai was used to conduct simulation experiments. The experimental results prove the effectiveness, robustness and advantages of the proposed algorithm.