台灣的營建產業中，預鑄構件的生產是屬於高度勞工密集且經驗導向的產業，本研究的研究目的是找出預鑄構件生產工項間的關聯法則，以得到工班調度和勞動成本最適化的結果。透過全面性的文獻回顧，本研究所收集的數據資料為全台灣近十年來的預鑄構件專案的生產時間，並花費數個月的時間建立資料庫，其中包括大樑、小樑、柱這三種主要的建築構件，每個構件包含14個生產工項，且分別有19558根大樑、16731根小樑以及18854根柱。本研究運用Pearson相關係數法確立工項和工項間的相關性，也運用了Apriori演算法，其中置信度的參數設定為90%，而支持度的參數設定為15%。本研究的第一個結論發現有四個生產工項並不需要高度技術工班，包含混凝土澆置、混凝土養護、脫模以及存放入庫，第二個結論發現，在大樑的工項中，放樣對其他生產工項有顯著性的影響，而在小樑和柱當中，則是預埋件埋設對其他工項有顯著性的影響，以上結論在工班調度以及勞動成本的最適化提供了明確的方向。;Manufacturing precast components in the Taiwan construction industry is a highly labor-intensive and experience-oriented process. The research objective is to explore the association rules among all precast activities in order that work scheduling and cost can be optimal. Gathering information from a comprehensive literature review, we aim at the entire precast production data for the data collection in Taiwan in the recent decade and it takes months to establish the database including three major components of pillar, large beam, and small beam. Each component contains 14 work activities and there are 18,854, 19,558, and 16,731 components for pillar, large beam, and small beam, respectively. We carry out the Pearson correlation analysis to determine the relationship among the activities, and adopt the Apriori algorithm with the parameter settings of confidence level at 90% and support value at 15%. The first finding indicates that 4 activities do not need to require high-skilled workers including pouring concrete, maintaining concrete, stripping module, and sending to inventory. The other finding points out that the layout activity significantly influences other activities for large beams; item pre-installation significantly influences other activities for both small beams and pillars. The findings provide a clear way for optimizing work shifts and labor costs.
