Adaptive & Discrete Real Bat Algorithms for Route Search Optimization of Graph Based Road Network

Bat Algorithm (BA) has recently emerged as an efficient nature inspired meta-heuristics due to its added parameters and searching features. In this paper modified versions of discrete bat algorithm is being proposed for the first time which will suit the discrete domain problems. Here the BA has utilized the famous three variable dependent Weibull Cumulative Distribution Function as Weibull Coded Binary Bat Algorithm (WCBBA), another as Real Bat Algorithm (RBA) and the third as the hybrid of the two, for search process with its modeling according to a road network management system where it is being tried to optimize the travel route and produce a vehicle load balancing structure of the network through optimized path establishment. In bat algorithm apart from the search criteria where the virtual bats move, they also utilize their Echolocation property for further investigation of the search space for prey. This makes the heuristic more probabilistic, dynamic and adaptive and the bats can reach a better solution through continuous analysis and exploration. Here the bat algorithm is modeled to handle a multi-objective optimization model where each bat tries to optimize its own route criteria or rather tries to find its liking prey. The enhanced search criteria of bats helps in reducing the number of bats for the search process, but as the Echolocation process spreads out it increases the complexity of the search process that is parallelism is traded off with complexity. The results show that the algorithm has potential for better results and has been compared with the converging rate of Ant Colony Optimization (ACO) & Intelligent Water Drops (IWD) algorithms.