The Unit for Systems Modelling and Analysis (USMA) is a research group comprising of post graduate students within the Department of Industrial Engineering at Stellenbosch University. The focus of the group is the research and application of advanced analytic methods. Research ideas and outputs are shared amongst members providing a platform for continual development.
Methods and techniques used include a combination of metaheuristics, computer simulation, mixed-integer linear programming, Markov chains, cooperative game theory and queuing theory. A large scope of the outputs fall under, though are not limited to, the field of Operations Research.
Much of the research within the unit deals with the field of discrete-event simulation (DES) in combination with metaheuristics as an optimisation tool. This applies in particular to dynamic, complex stochastic processes where the solution space is discrete and the number of possible solutions is too large for a formal experimental design. Also there might exist multiple, conflicting objectives. Enter multi-objective optimisation, the rich field of research that the group spends much effort upon. Metaheuristics such as the genetic algorithm (GA), population-based incremental learning algorithm (PBIL) as well as the Harmonic Search algorithm (HS) are used in simulation optimisation. The cross-entropy method, tabu search and simulated annealing are also used as optimisers. One aspect of the research is to find more efficient and faster algorithms – by modifying them or developing new ones – to mitigate the computational burden imposed by simulation models of real-world systems while ensuring statistical confidence. Industrial problems provide engineering applications and validation test benches for optimisation models.
The SUnORE (acronym for Stellenbosch Unit for Operations Research and Engineering) group focusses on the development of mathematical models and their incorporation into computerised systems aimed at supporting scientifically justifiable and effective decisions in industry. These models draw from the scientific fields of applied mathematics, statistics engineering and computer science, and are applicable in the context of complex problems which admit a large variety of trade-off solutions. The group has forged strong decision support ties with a number of industry partners in the agricultural, retail, banking, insurance and military sectors. The group also collaborates with and supports a number of parastatals, NGOs and non-profit organisations. Examples of the types of decisions for which decision support is typically provided by the group includes (but is not limited to):
In addition to producing work of a high standard, the group also strives to create an inclusive environment for the training, development and inspiration of innovative, yet responsible, scientific young minds.
You may visit the SUnORE website by clicking here.