Product Portfolio Management with Discrete-Event Simulation and Genetic Algorithm

Abstract

This paper presents a discrete-event simulation environment that models the underlying structure of a pharmaceutical enterprise portfolio. An object-oriented model structure previously developed for batch plant scheduling and design is extended to embed the case of product management, which is particularly adequate for re-use of both structure and logic. By wrapping such a simulation with advanced optimization approaches, such as Genetic Algorithms, it becomes possible to evaluate a large set of scenarios for the pharmaceutical enterprise, with a holistic fashion that avoids local optima that can be detrimental to the enterprise.