營建業多能外籍勞工指派最適化

Abstract

面對外勞技術能力的不一致性，管理者通常以直覺或印象來評估每位外籍勞工的能力特質，並根據其技術特性進行指派工作。然隨著工作面與同步作業的增加，具有某些專長的外勞將可能成為限制性資源；某部份的外勞卻因不太擅長該作業或缺乏該作業專長而產生閒置。管理者在面臨外籍勞工與本地技術工工資的差異與趕工加班所招致之額外成本，甚至須考量工作彈性調度的結果所造成之生產力折減，當各項限制條件複雜時，管理者將對於外籍勞工指派及外聘本地技術工人數之模式難以果斷且明確地決策，導致工程人力成本無法有效控制。很明顯地，在分派外勞人力的過程中，缺乏一種具有合理且優化的分派機制。 本研究嘗試以整數規劃針對此問題建立一系統且合理的外勞指派數學模式，在有限的人力下，同時考慮外籍勞工各項生產力之不一致性以及彈性調派所招致之成本，決定整體施工團隊生產力最高且相對成本最低的分派模式。在求解方面，本研究採用遺傳演算法作為模式求解工具，輸入各項參數與限制條件，以求得排班天數內外籍勞工最適當的工作指派，以及當需求工作量超過外勞工班所能負荷時，提供尋求本地技術工支援之數量與類型之決策參考。

The foreign workers were usually assigned works according to their productivity assessed with the managers’ intuition or impression, when facing the inconsistency of foreign worker's productivity. By the way, the foreign workers with some special skills may become restricted resources, and some foreign workers maybe idle because of lacking some special skills with the increase of parallel operations and work faces. Managers must consider the different salary among the different kind of local technicians and the cost caused of the foreign workers overtime working. Furthermore, managers also have to think over the reduction of productivity due to the flexible assignment. As every limiting condition complicated, It is difficult for managers to make decision resolutly and clearly. Obviously, it seems to lack a kind of rational and systematical mechanism during the process of assignment. This study develops an integer programming model that consider the cost caused in inconsistency productivity of each foreign worker and the reduction of productivity due to the flexible assignment to facilitate manpower directly on the most economy assignment model. After the model was set up, Genetic Algorithm was applied for searching the optimum solution. The solution calculated by Genetic Algorithm can assistant managers to assign the foreign workers to the suitable tasks that appropriate for the lowest manpower expenditure.