A Hybrid Code Genetic Algorithm for VRP in Public-Private Emergency Collaborations
Shaoren Wang, Qiming Huang
Abstract
To minimize the total time for the distribution of relief commodities participated by both private companies and the government, a vehicle routing problem (VRP) model in emergencies was proposed. Considering the differences in the starting points of vehicles, the VRP of general logistics, and departments of vehicles, constraints, such as vehicle capacity limitation and time windows, were introduced into the model, which was close to meeting the practical demands of emergency relief. A hybrid code genetic algorithm (HCGA) was proposed, and it used dynamic mutations to avoid early traps in local optimization and to accelerate convergence. This algorithm was programmed by MATLAB. Furthermore, the vehicle routing optimization plans in an emergency was calculated by a simple genetic algorithm (SGA) and the HCGA, respectively. Results demonstrate that the total time for relief distribution in the HCGA is 11.62 % lower and the calculation time is 14.24 % shorter than that of the SGA. The HCGA is not only convenient in processing the constraints of the model and the natural description of problem solutions, but it is also effective in improving the complexity.