Research On Muti-objective Emergency Scheduling Model And Improved Differential Evolution Algorithm

In recent years,major accidents and natural disasters have occurred frequently in China and around the world which Resulting in huge casualties and economic losses.Therefore,it is really important to build up an emergency solving management system which can have a rapid response and a sensitive early warning system to give the reasonable emergency resource rescue method.Good strategy will increase the delivery efficiency and decrease the further damage afterwards.So the emergency resources scheduling problem research has been widely attention.Because the emergency scheduling problem is a typical multi-objective optimization problem,multiple goals often limit each other,the traditional solution is difficult to converge to the Pareto solution set.Therefore,the emergency rescue model is established and the algorithm is solved.The two aspects of emergency response are discussed and studied.For the mathematical model,the multi-objective emergency resources scheduling model of the multi-supply points less than the multi-disaster points in the case of limited resources is established for the practical demand of the emergency resources scheduling problem.For the algorithm solving,evolutionary algorithm is improved,and the concept of Pareto and crowding distance are applied to solve the multi-objective optimization problem.A new DE algorithm search equation is designed to study the optimized DE capability in solving the continuous optimization problems.The specific research content is as follows:Firstly,the assumption background is based on the nonlinear continuous consumption emergency resources scheduling.It includes the network between supply points to the disaster points and also considers the relationship between the limited delivery capabilities and the huge amount of goods required.Therefore,the multi-objective and multi-limitation emergency resources scheduling model is created with an aim that minimizes the total delivery cost and the value of top loss as well.Secondly,In the case of the basic DE algorithm,the performance of the population unit is evaluated according to the concept of Pareto domination and crowding distance.For the non-dominated solution at the same level,Pareto sorting is carried out by using the crowding distance concept.Meanwhile In order to increase the capability of the algorithm,DE/best/1 evolution strategy and DE/rand/2 evolution strategy have been involved and they are combined together to a new double-evolution strategy This paper involves DE/best/1 mutation strategy and DE/rand/2 mutation strategy.For the population units with Pareto frontier solution,it's good to use the DE/best/1 mutation strategy because the DE/best/1 has good development ability.For the population units with non-Pareto frontier solution,it's better to use the DE/rand/2 mutation strategy because it has good exploration ability.Finally,from the simulation experiment,it proves that under the aim with minimum of total delivery cost and cost of top loss,compared with basic differential evolution algorithm,the improved differential evolution algorithm achieves better result to solve the multi-objective and multi-limitation emergency resources scheduling problem.The improved differential evolution algorithm also has better balance between the development ability and exploration capability.It also increases the volume of the Pareto solutions,improves the convergence of solution possibility and makes the solutions in the population widely.These prove this paper gives a reliable emergency resource scheduling and an available improved differential evolution algorithm.