應用NSGA-II和TOPSIS在生產多種零件下可重構製造系統最佳化的規劃

OPTIMAL CONFIGURATION PLANNING FOR RECONFIGURABLE MANUFACTURING SYSTEMS WITH MULTIPLE TYPES OF PARTS USING NSGA-II AND TOPSIS

以可重構工具機為基礎的可重構製造系統，因可重組機台組態以因應生產產品的變動，所以成為製造系統發展的重點。過去的研究著重以生產單一產品單序列流程的可重構製造系統的機台組態及數量的決策為主，然而，可生產多種產品的可重構製造系統，不但可以減少機台組態的重組，更能快速回應產品變動及多樣性的需求。不同的產品有不同的生產階段及作業，規劃時需要整合不同產品的生產階段及作業，以決定機台組態及數量。而且多種產品共同生產時，可能共用一台機台來生產，作業流程不再是單一個序列，而是多個序列組成的非對稱配置，讓規劃問題變的更複雜。本研究在考慮滿足各種產品作業需求及產能需求下，以成本、可重構能力、操作性能、擴展性四個目標指標下找出生產多種產品的可重構製造系統最佳的配置，系統的配置包括階段數、階段間的連結方式、各階段作業、作業所使用的機台組態及其數量。最後，以結合非支配排序基因演算法II及逼近理想解排序法在四個目標指標下找出最佳的系統配置。

The development of a reconfigurable manufacturing system (RMS) based on reconfigurable machine tools (RMTs) becomes an important issue, for its capability of reassembling machine configurations responds to product changes. Previous researches focused on configuration selection and number determination of machines for RMSs, producing a product with a single sequence of processes. However, an RMS, capable to produce a variety of products, not only reduces reassembly of machine configuration, but also quickly responds to the needs of product diversities and changes. Different products have different stages of production and operations so it is necessary to integrate these different stages and operations for determining the number and configuration of machines. Moreover, when a variety of products are produced together, machine sharing may occur, and the process flow is no longer a single sequence, but an asymmetric configuration of multiple sequences. Thus, planning of the RMS becomes more complex. In this research, a multiple-objective optimal configuration selection for the RMS, producing a variety of products, is proposed. The four objectives considered are cost, machine reconfigurability, operational capability and scalability. The configuration selection includes the number of stages, connection type between stages, operation in stages, the number and configuration of machines. Finally, non-dominated sorting genetic algorithm II (NSGA-II) and technique for order preference by similarity to ideal solution (TOPSIS) are applied to find optimal configuration selection of the RMS.