Research Article
Machine Learning-Based Parameter Tuned Genetic Algorithm for Energy Minimizing Vehicle Routing Problem
Algorithm 1
Overview of basic genetic algorithm developed for the EMVRP.
Input | Read list of cities to be served with demands from CVRP Lib file | Read vehicle capacity from text file | Population size () | Number of generations () | //Initialization | Create a tour with a random order of cities | Do this for times to create a population | Get the best tour from the population | Save as the elite | initial energy = energy consumption of the fittest tour in first population | //Genetic algorithm | //Run for times | Loop1 | //Run for times | Loop2 | //Tournament selection | Select a random set of tours from the population | Get the fittest and return | //Crossover | Parent1 = tournament selection () | Parent2 = tournament selection () | child = Crossover (Parent1, Parent2) | //Mutation | Swap random two cities in the child | endLoop2 | //New population is created | Get the fittest | Replace previous elite if fittest is better than elite | endLoop1 | Get the elite | Final energy = energy consumption of elite | Reduction of energy = (initial energy − final energy)/initial energy 100% | Print “Reduction of energy” |
|