Improved Differential Evolution Algorithm to Solve Bi-objective Emergency Material Scheduling Problem

We aim to establish a suitable emergency material scheduling model and optimize a differential evolution algorithm to solve the model. Under the condition of limited resources, this model aims to obtain the minimum total cost of delivery and the maximum reduction of disaster sites about continuous consumption, considering multiple supply points and multiple disaster points. According to the concept of Pareto domination and crowding distance, differential evolution algorithm has been used to solve constraint and bi-objective models. We optimize a differential evolution algorithm by combining DE/best/1 variation strategy and DE/rand/2 variation strategy to form an improved algorithm involving a double-mutation strategy, which increases the algorithm's scapability. Our simulation test proves that the model and the algorithm are feasible and effective. According to comparison results involving big and small scale examples, the proposed algorithm can decrease the total delivery cost of emergency resource scheduling and obtain the maximum reduction of the disaster sites. Meanwhile, this improved algorithm increases the quantity of the Pareto non-dominated front solutions and increases extension distribution of the solutions.