Development and application of a hybrid genetic algorithm for resource optimization and management

Resource selection/optimization problems are often characterized by two related problems: numerical function and combinatorial optimization. Although techniques ranging from classical mathematical programming to knowledge-based expert systems (KBESs) have been applied to solve the function optimization problem, there still exists the need for improved solution techniques in solving the combinatorial optimization. This paper reports an exploratory work that investigates the integration of genetic algorithms (GAs) with organizational databases to solve the combinatorial problem in resource optimization and management. The solution strategy involved using two levels of knowledge (declarative and procedural) to address the problems of numerical function, and combinatorial optimization of resources. The research shows that GAs can be effectively integrated into the evolving decision support systems (DSSs) for resource optimization and management, and that integrating a hybrid GA that incorporates resource economic and productivity factors, would facilitate the development of a more robust DSS. This helps to overcome the major limitations of current optimization techniques such as linear programming and monolithic techniques such as the KBES. The results also highlighted that GA exhibits the chaotic characteristics that are often observed in other complex non-linear dynamic systems. The empirical results are discussed, and some recommendations given on how to achieve improved results in adapting GAs for decision support in the architecture, engineering and construction (AEC) sector.