An analysis of Artificial Immune System and Genetic Algorithm in urban path planning

Evolutionary Algorithms like Genetic Algorithm (GA) and Artificial Immune System (AIS) are commonly used to find solutions to problems not suitable for traditional optimization approaches. In this study, we compare the results of AIS and GA for path-planning where the objective is to optimize the safety and the travelling distance. Since these algorithms are computationally intensive, we perform offline optimization to generate a list of suboptimal solutions. Results show that the performance of GA and AIS are similar in terms of convergence and optimality. Furthermore, an analysis of AIS revealed that the convergence rate is faster at higher separation threshold; however, the effects of maturity age and percentage of hypermutation had minimal effects in convergence of AIS. Using AIS, we were also able to produce several sub-optimal paths in the form of memory cells, which provide robustness to the optimal path subject to perturbations.