Empty flow analysis and compact load transport networks in unequal area facility layout problem
研究了不等面积设施布局问题中，如何基于启发式算法输出设计物料运输网络，以优化装载运输与空流的总和，并发现空流会显著增加运输距离，而紧凑网络可大幅减少所需网格长度。
The unequal area facility layout problem (UA-FLP) involves determining the placement of unequal-sized workcenters within a facility to minimise total loaded transport distances. The load transport network is defined as the grids formed by a set of horizontal and vertical lines connecting the centroids of the workcenters (c2cHV lines). UA-FLP heuristics typically report a marginal improvement of 1%–5% in loaded transport compared to previous approaches. In this study, we design material transport networks based on the outputs of UA-FLP heuristics to optimise the combined loaded transport and empty flows. Two dispatching policies, shortest-trip-distance-first (STDF) and first-come-first-served (FCFS) are implemented to account for empty flows. For benchmark problems, we show that, on average, under the STDF and FCFS policies, respectively, 71% and 152% empty flows are added to the best solutions obtained by UA-FLP heuristics for loaded transports. Furthermore, due to inherent symmetries in UA-FLP designs, where all workcenters are rectangles, numerous optimal solutions exist for both loaded transport minimisation and combined loaded transport and empty flow minimisation. Identifying the minimal-length network that achieves optimal performance for both metrics is another design question addressed in this study. For our benchmark problems, the minimal-length network required to achieve optimal performance, on average, uses only 23% and 31% of the total c2cHV grid length under the STDF and FCFS policies, respectively. This study provides additional metrics for UA-FLP designers to assess the quality of their layouts and compare them with previous designs based on not only loaded transport but also total loaded transport and empty flows, as well as the minimal network length required to achieve optimal loaded and empty flows.