| The frequent occurrence of large-scale disasters has brought great impact on the normal lives of people,and persons and property suffered extremely heavy losses around the victims.In some cases,there is no sign of disaster therefore,emergency management departments should be ready for emergency rescue at any time.The timely and effective emergency rescue schemes should be designed,and the highly efficient rescue actions should be operated,so that the loss of people in the affected areas can be mitigated as much as possible.The emergency resource scheduling(ERS)is an important section of the emergency rescue,and it is intended to reasonably allocate the limited emergency resources to the various affected areas,to meet the post-disaster demands of affected areas.In this paper,based on the multi-period dynamic ERS problems,the reasonable scheduling models are built,the corresponding evolutionary algorithms are put forward to solve these models,and reasonable and time-saving emergency scheduling schemes are obtained eventually.The main works are summarized as follows:(1)For the multi-period dynamic ERS problem with the uncertain traffic network,a single-objective model for multi-period dynamic ERS problems is built,which takes the weight sum of the unsatisfied demand of disaster areas,the risk of transporting resource and the vehicle cost during the scheduling as the objective,and a revised multi-agent genetic algorithm(labeled as MAGA-md ERS)was proposed based on the intrinsic properties of the problem to solve the model.Beside,we designed the appropriate crossover and mutation operators,and local search and elitism strategy are also applied in our algorithm.The performance of genetic algorithm(GA)and MAGA-md ERS were compared through lots of contrast experiments,and the effectiveness of the proposed algorithm was verified.The applicability of the proposed algorithm in large-scale disasters was also tested with three large-scale instances.(2)In order to minimize the unsatisfied demand of affected areas and the risk of transporting resource,a multi-objective evolutionary algorithm based on decomposition for multi-period dynamic emergency resource scheduling problems(MOEA/D-md ERS)was proposed,and the simulated binary crossover(SBX)operator and the real mutation operator based on adjustment with the intrinsic properties of the problem were designed in the algorithm.The effectiveness and the applicability of MOEA/D-md ERS in large-scale instances were verified through plenty of experiments.Besides,the performance of MOEA/D-md ERS and the non-dominated sorting genetic algorithm II(NSGA-II)on the proposed model were compared,and the experimental results showed that MOEA/D-md ERS can provide the more efficient and time-saving scheduling schemes for the multi-period dynamic emergency resource schedulings.(3)Considering the dynamic of the demand of affected areas in the process of scheduling,we built a multi-period dynamic emergency resources scheduling model under dynamic demand,and put forward a specific multi-objective evolutionary algorithm based on decomposition with differential evolution(labeled as MOEA/D-DE-md ERS)to solve the model.In this algorithm,three adjustment strategies were designed to dynamically adjust the transport schemes during the scheduling process,and meet the variational demands of people in affected areas as much as possible.The way of adjustment of scheduling schemes was presented in a small-scale instance when the demand changes dynamically.The contract experiments betweent MOEA/D-DE-md ERS and the multi-objective evolutionary algorithm based on decomposition with the simulated binary crossover(MOEA/D-SBX)showed the effectiveness of the proposed algorithm in the multi-period dynamic emergency resource scheduling problems under dynamic demand. |
PDF Full Text Request