A demand-to-manufacturing-requirement model based on predefined configurations

This paper examines the production planning of a job-shop manufacturing system capable of producing various part types. The demand for the various product types is known at the end of each elementary period. The objective function is the sum of inventory and backlog costs accumulated during a finite period of time known as the horizon of the problem. This horizon is a multiple of the length of an elementary period. The authors define the new concept of configuration as a set of maximal quantities of each type of parts which can be manufactured during an elementary period of time. This set is chosen either in order to balance the resource loading or to optimize the use of the tools available in the job-shop. The authors assume that there exists a finite set of configurations. The changeover of the system from one configuration to another is only allowed at the end of an elementary period and introduces a set-up time during which no machine can produce. The control of such a system consists in choosing the sequence of configurations and the actual production levels on each elementary period so as to minimize the objective function. The authors propose an efficient optimal algorithm for computing the actual production levels on each elementary period when the sequence of configurations (i.e. the sequence of constraints) is given. A heuristic algorithm is also proposed to find a good sequence of configurations.