Place: Large Lecture Room
Affiliation: Department of Computer Science, West University of Timisoara, Romania
Differential Evolution (DE) is currently one of the most used population based stochastic metaheuristic for solving multi-modal, multi-objective, dynamic or constrained optimization problems. Despite its simplicity in implementation, DE is a search method with a complex behavior which is not yet fully understood, being difficult to give answers to questions like “Why is DE successful for a class of problems and why does it fail for other problems?” or “How could we control the DE behavior?”. The theoretical analysis of DE is still well behind the experimental results, most of the current knowledge on differential evolution being based on empirical observations.
This presentation will first provide an overview of the main DE variants and on existing theoretical results, focusing on those which provide practical insights which are useful in designing DE algorithms which can avoid undesirable behaviour (e.g. premature convergence). The second part of the presentation will shortly review several DE applications in parameter estimation (airfoil design and thymocyte dynamics simulation), in data mining (clustering and rules mining) and image processing. Finally, some scalability issues will be addressed and particularities of two parallel implementation models (master-slave and island model) will be discussed.