考虑同时包含单处理机和批处理机的多阶段混合流水车间调度问题，设计一种基于时间窗的蚁群算法，以最小化最大完工时间为优化目标。在解决整体调度问题的过程中，通过蚁群算法实现工序分派、排序、组批三个阶段的协同优化调度;另外在工件组批阶段加入时间窗策略，利用前瞻性算法动态构建组批方案，通过分析批处理机的时空关系设计合理的组批时间窗，以获得较高的机器利用率。试验结果表明，与无时间窗的蚁群算法相比，时间窗策略在确保最小化最大完工时间的同时，提高了批处理机的利用率;与CPLEX相比，基于时间窗的蚁群算法在最大完工时间和计算效率方面均有较大优势。蚁群算法与时间窗的结合兼顾了多个阶段多种机型的生产特点，适用于解决多阶段混合流水车间的调度问题。

The scheduling problem of a multi-stage hybrid flow shop (HFS) with single processing machines and batch processing machines is addressed. Aiming at minimization of the maximum completion time,a time window-based ant colony optimization (TWACO) approach is proposed,which achieves cooperative optimization for part assignment,part sequencing,and batch formation simultaneously. Besides,to obtain higher utilization rate for the batch processing machines,a time window is dynamically constructed by a look-ahead algorithm during batch formation. The computational results show that by the adoption of the look-ahead algorithm,TWACO maintains a good balance between the utilization rate of the batch processing machines and the solution quality; and TWACO has significant advantages over CPLEX in minimizing the maximum completion time while retaining low computational costs. Therefore,TWACO is much more suitable for the problem of a multi-stage HFS.