Generic parallel genetic algorithms framework for optimizing intelligent transportation systems (GENOTRANS)

Many intelligent transportation systems (ITS) require formulations of Optimization problems. Heuristic solution methods such as genetic algorithm (GA) are successful and attractive approaches to solve such optimization problems. Additionally, ITS optimization problems are computationally demanding and often intractable by sequential GA.;To demonstrate the use of the new tool, GENOTRANS was tested on the optimization of GAID (Genetic Adaptive Incident Detection) which utilizes multiple smoothing parameter probabilistic neural networks optimized by genetic algorithms. The Master-slave parallel scheme has shown a significant speed up relative to sequential GA. The multi-deme approach generally enhanced the quality of the solution and the convergence characteristics.;This thesis introduces the architecture and development of a generic distributed parallel genetic algorithms engine (GENOTRANS). It utilizes parallel population structures that proved to enhance the GA overall performance. The presented framework is based on the Java 2 Enterprise Edition (J2EE) architecture that provides standards for building enterprise solutions.