Two heuristics for the capacitated minimum spanning tree problem with time windows
针对带时间窗的容量限制最小生成树问题，提出两种启发式算法，在测试集上优于Solomon的经典方法，且计算效率高。
We consider the capacitated minimum spanning tree problem with time windows (CMSTPTW) and propose an enhancement of greedy heuristic for the uncapacitated version of it, showing that our algorithm significantly outperforms Solomon’s one based on the R7 test problem. To examine the performance of our approach, we convert the vehicle routing data sets to appropriately model CMSTPTW instances and provide solutions for all of them; in addition, we compare the solutions we reach for the uncapacitated version of all instances and those obtained by Solomon to demonstrate the consistent superiority of the proposed method. Finally, we develop a second solution approach, the local domain priority heuristic, which when applied to the same data sets results in better solutions for the majority of the CMSTPTW instances, without excessive computational effort.