Emergency Logistics Network with Public–Private Cooperation and Multiple Deliveries: An Optimization Approach

An important way to cope with a large-scale disaster is through relief activities coordinated by public and private enterprises engaged in emergency responses. However, the distribution of relief supplies in a disaster area is often chaotic because emergency logistical systems typically lack a program with a collaborative public–private facility and optimization of vehicle routing. To solve this problem, we propose the creation of a method for optimizing the locations of facilities and the paths of vehicles for public–private cooperation after a disaster, including selecting potential centers for distributing emergency supplies, dispatching vehicles from different parking lots to distribution centers to load supplies, and optimizing vehicle routes to deliver relief supplies to disaster areas as soon as possible. In this way, secondary distribution routes for relief supplies can be continuously optimized. In particular, this study examines decision making on the problem of designating facility locations, the vehicle-routing problem, and the follow-up problem of routing vehicles to multiple distribution centers. After determining the characteristics of the optimization models for solving these problems, we propose an improved genetic algorithm (GA) based on an encoding and field restriction operator, a hybrid ant colony optimization algorithm, and a hybrid particle swarm optimization algorithm. This improved GA achieves a natural representation of solutions to the aforementioned problems, thereby facilitating the treatment of constraints. Moreover, we mathematically deduce the time complexity of the objective function evaluation of the improved GA. Lastly, we confirm the effectiveness and efficiency of the models and algorithms through a numerical case study.