資源限制專案排程問題,長久以來一直是個相當複雜而困難的組合最佳化問題,一般而言,此類的問題常受到活動執行順序及資源使用量的雙重限制,它的特性是求解所需花費的時間會隨著問題複雜度大幅的增加。本研究建構一資源限制專案排程模式,探討在不同的專案作業網路及資源需求環境下,7種作業優先規則之最小化總專案工期績效表現。總結實驗結果,本研究提出之作業優先規則LRD/LFT ,因兼顧時間及資源因素,得到或並列最佳解次數最多,比排名第二僅考慮時間之作業優先規則LFT高出9%之多。然而,並沒有一種作業優先規則可以求解各式題型的問題,研究整理後發現,依總資源使用密度,作業優先規則的選擇是有規則可尋的。我們提出綜合性作業優先規則,當總資源使用密度約30%以下,取α=0,β=1時,亦即作業優先規則為LFT時,可得較優總專案工期績效;當總資源使用密度約在30%以上,取α=1,β=1時,亦即本研究的作業優先規則LRD/LFT作業優先規則,可得較優總專案工期績效;而當總資源使用密度約在60%以上且資源使用種類為單資源時,取α=1,β=0時,亦即本研究提出的另一作業優先規則LRD,可得較優總專案工期績效。這是先前研究中,所未曾提出的優先法則選用規則。
Resource-Constrained Project Scheduling has been known as NP-complete combinatorial problem. In this study, we compare 7 kinds of priority rules, including two new rules we proposed, under different project networks and resource requirements. The experiment results show that the frequency of the minimum project completion times through applying the priority rule LRD/LFT , which considering both factors of task duration and resource requirements, is more than that of the rank second rule LFT by 9%, which only considering the factor of task duration. However, it is well recognized that no priority rule can get the best solutions in all different settings and objectives. We come up with the synthetic priority rule , which can be systematical applied in different project environments according to different resource requirement density. When the density of the total resources usage is under 30%, we suggest α=0,β=1 to let the rule become the LFT rule. When the density of the total resources usage is more than 30%, we suggest α=1,β=1 to let the rule become LRD/LFT priority rule we proposed. The density of the total resources usage is about more than 60% and only single resource is required, we suggest α=1,β=0 to let the rule become LRD rule which is a simple rule we proposed. The selection criteria among the priority rules has never been brought up in previously researches.