Custom scheduling algorithms in a simulation-like environment

We present a new framework that allows users to easily develop scheduling algorithms that are both highly customized and powerful. This environment is a unique and simple way to satisfactorily solve real-world problems using embedded Problem Space Search heuristics, Genetic Algorithms, and user-coded models. We use an easy and powerful methodology that utilizes simulation-like constructs borrowed from the discrete event simulation field for developing the environment, since simulation is a straightforward way to describe a specific manufacturing system. We show that this solution methodology is easily adapted to various manufacturing settings and it produces high-fidelity shop floor models. We illustrate this idea using real industrial shop floor scheduling problems that contain simultaneously: lot sizing, batch aggregation and disaggregation, alternate routings, sequence dependent setup times, and due date constraints. Problems from three different industries are tested in order to prove real-world applicability and demonstrate robustness. We also solve jobshop problems from academia and introduce modified jobshop problems with alternate routes. The ease of a simulation-type language is demonstrated. We argue that our framework overcomes shortcomings of currently available methods that can be too general, too costly, and/or too simple.