A note on the tool switching problem of a flexible machine

The problem of minimizing the total number of tool switches for a numerically controlled flexible machine is considered. A set of parts is to be processed with the machine. Each part needs a set of tools which should reside in the magazine of the machine at the moment of processing. Because of the limited capacity of the magazine, tools must be switched and the objective is to minimize the amount of this work. We propose an algorithm which tries to avoid sticking to a local minimum by repeated searches from different initial starting points which are created by repeated construction of super parts from parts with similar tools. The proposed algorithm and a number of efficient heuristics presented in the literature are empirically tested by both random test problems and real production data. The new algorithm performs well when considering the tradeoff between solution quality and running time.