Applications of GA and Hybrid GA in various domains: Design and Prospects

One of the most widely adapted algorithms in bioinspired optimization techniques is the Genetic Algorithm (GA). It has been used extensively in solving problems that require a Pareto-optimal solution for bi-objective optimization problems. The general working of GA followed by diverse applications, particularly in resource scheduling is exemplified in this work. The application of GA in flexible job shop scheduling problems (FJSSP), University Course scheduling, the role of GA in preventive maintenance, automated manufacturing systems, cross-deck scheduling, and conflict-free scheduling of Automatic Guided Vehicles (AGV) are studied. The formulation of the fitness function and various strategies used to improve the existing crossover or mutation are examined. Some heuristics that are inspired by GA and the effect of combining GA with other heuristics are also reported. GA with ANN, GA with Ant-Colony Optimization, Multi-objective GA (MOGA), and GA with random forest learning algorithm are explored. The chromosome representation, the various selection, crossover, and mutation strategies are compared. The hybrid versions of GA show an improvement in the quality of solutions and comparatively lesser deviation from the optimal solutions. The increasing time complexity is a demerit in such versions and there is scope for further work in proposing a better algorithm with lower time complexity. The merits and shortcomings of GA and the scope for further research in this area are put forth.