本研究針對完全相同平行機(Identical Parallel Machines)允許中斷(Preemption)條件下,探討總延遲時間最小化(Minimizing Total Tardiness)排程問題,此排程問題可表示成Pm|pmtn|∑Tj ,此問題為NP-hard問題。 本研究利用分枝界限法(Branch-and Bound Algorithm)求解所探討的排程問題,以派工法則決定工件的完工順序,並根據Prot[12]所提出的線性規劃(Linear Program)求解所對應的PFS-like排程進而得到起始上限值,接著根據Schaller[7]提出的下限值演算法(LB4),針對每一節點所代表之次問題求解下限值,並在分枝的過程中,刪除不必要的節點,增加求解的效率。 本研究探討完全相同平行機總延遲時間最小化排程問題,並透過電腦隨機產生大量的測試例題評估所發展上限值、下限值及整體分枝界限法之求解績效,並分析控制到期日寬鬆程度(τ)與控制到期日範圍(R)這兩個參數對問題困難度及求解效能的影響。 本研究之上限值演算法與最佳解之平均誤差為4.01%,下限值演算法與最佳解之平均誤差為41.55%,本研究所發展之分枝界限法可在有限時間內正確求解16個工件及5部機器的問題。
This study examines the problem of scheduling preemptive jobs on identical parallel machines to minimize total tardiness.The problem can be expressed as Pm|pmtn|∑Tj , and is known to be NP-hard. An efficient heuristic is developed for solving large sized problems. In this heuristic, a job completion sequence is first determined by a dispatching rule, and then the linear program from Prot[12] is used to find the corresponding PFS-like schedule. A lower bound scheme from Schaller[7] is also presented. Both of the proposed heuristic and lower bound are incorporated into a branch-and-bound algorithm to optimally solve small sized problems. Computational results are reported. The branch-and-bound algorithm can handle problems of up to 16 jobs and 5 machines in size within a reasonable amount of time. The solution obtained by the heuristic has an average percentage deviation of 4.01% from the optimal value, while the initial lower bound has an average percentage deviation of 41.55% from the optimal value. Moreover, the heuristic finds approved optimal solutions for more than 45% of the problem instances tested.