An Improved Genetic Algorithm Based Solution to Vehicle Routing Problem over OpenMP with Load Consideration

Vehicle Routing Problem being a combinatorial class of problems has implications in various areas and applications where traditional methods to find a search space either fail or slow down especially in case of real time systems. Using the most popular heuristic searching approach, Genetic Algorithm, there have been solutions in the state of the art to the problems like Vehicle Routing, Traveling Sales Person, etc. In this paper, a parallelized version of genetic algorithm has been proposed for vehicle routing problem over OpenMP programming model. The problem has been solved taking into consideration the constraint that congested cities of the network have been pushed behind after obtaining a fit set of chromosomes in every iteration so that as the time pass by the load sheds on the cities and we optimize in terms of time as well. The concept can very well be mapped to the network routing problem as well. Experimental results show that the efficiency of the proposed model is high in comparison to the serial version where the machine certainly meets the restrictions in terms of execution and processing power to get the intensive computations done as the number of cities or nodes in the network increases. Finally, the paper sums up with the future endeavors of the problem with more complex constraints getting involved to fetch the best possible route and how parallel processing must deliver such solutions.