In this paper, the authors propose a Case-Based Reasoning (CBR) approach for non-preemptive scheduling of independent jobs n to unrelated parallel machines m where the time required for the processing of a job on a machine is a function of both the job and the machine, and the objective is to minimize the makespan. The CBR approach makes direct use of previous cases (problems and their solutions) and finds near-optimal solutions based on similarities between problems and cases. The present discussion is focused on investigating the efficiency of the CBR approach corresponding to the degree of similarity between problems and cases. Computational experiments show that this approach can find very promising solutions in very short running times even if problems are not very similar to cases and so it is an efficient and effective approach. [In Japanese.]
