Method able to find an optimal solution to an optimization problem. Such method are not appropriate for a NP-hard problem, except if its size (e.g., number of decision variables) is small.
Published in Chapter:
A Simulation-Optimization Approach for the Production of Components for a Pharmaceutical Company
Nicolas Zufferey (University of Geneva, Switzerland), David Dal Molin (IPros, Switzerland), Rémy Glardon (IPros, Switzerland), and Christos Tsagkalidis (IPros, Switzerland)
Copyright: © 2018
|Pages: 15
DOI: 10.4018/978-1-5225-2944-6.ch013
Abstract
The considered problem (P) concerns the production of strains (also called jobs or batches), which are the used components in the final products that are bought by the consumers. (P) contains two components that have to be tackled sequentially: the inventory management problem (IMP) and the job scheduling problem (JSP). (IMP) is solved with a reorder-point policy, defined on the basis of critical demand coverage. To tackle (JSP), a descent local search (DLS) is used, based on swap moves. In other words, for a given job sequence, a series of modifications is performed on it in order to try to improve the solution, where each modification consists of exchanging the positions of two jobs. Because of random events (some jobs might be rejected if they do not meet predefined standards) and stochasticity (the duration of each job follows a normal distribution), simulation is required to evaluate any sequence of jobs that is a solution to (JSP). A simulation-optimization approach is therefore proposed to accurately tackle (JSP). This work is motivated by a real pharmaceutical company.