A Real-world-Like Evolutionary Algorithm on the Cloud-Computing Environment Ming-Shen JIANa, Ta-Yuan CHOUb aDepartment of Computer Science and Information Engineering, National Formosa University, Yunlin, Taiwan b Department of Computer Science and Engineering, National Sun Vat-sen University, Kaohsiung, Taiwan Corresponding Author: Ta-Yuan Chou, tayuan@gmail.com Abstract-This paper develops a multiobjective evolutionary algorithm on the cloud computing environment to help planners solve multiobjective problems more efficiently and effectively. The cloud environment is emulated as a virtualized biological world with several isolated regions. The main population initially continues evolutionary processes as the most widely-known evolutionary algorithms. To yield both exploration and exploitation, two processes, such as migration and interaction, are deployed. In the process of migration, local optimal solutions can migrate to form new populations so that the search space can be expanded. To overcome the disadvantages in isolated evolutionary algorithms, the individuals in different populations will interact stochastically in the interaction process. Taking the advantage of cloud computing environment, planners can take less effort on deploying both computation power and storage space. Instead, the planners can focus on the design of encoding, crossover, and mutation. Also, it can further applied in various complicated applications more practically. Kqword-Evolutionary Computation, Cloud Computing I. INTRODUCTION Emergency logistic is extremely important, when natural disasters, such as toadoes, typhoons, tsunami, earthquakes, and re teor happen more equently. When natural disaster occurs, most of these places will require a large quantity of resources, such as food, water, medical supplies, tents and clothes. However, these damages may destroy roads too. Under the circumstances, some places will be isolated. How to deliver required resources to these places as soon as possible becomes tedious. Traditionally, the emergency logistic can be categorized as several aspects. Some researches focus on road reconstruction. Some cluster-based researches focus on nding clustering node. Also, some researches focus on routing selection. general, the problem can be viewed as one type of vehicle routing problem (VRP). Since the VRP can be rther categorized as VRP with pickup and delivering, VRP with LIFO, VRP with time windows (VRPTW), and capacitate VRP without or with time windows (CVRP or CVRPTW), the factors of the vehicle routing problems are time limitation, capacities, and path selection. Assume there is a group of commodities, the emergency logistic problem is to nd out the vehicles and the routes to deliver the supplies. The feature of natural disasters is emergency. Natural disasters will cause the broken of transportation inastructure, water and power supplies. Also, the natural disasters will cause the damages of lives, shortages of materials, broken of houses. When these damages occur, the rescuing processes should complete the tasks within the golden 72 hours. Furthermore, some diseases will occur aſter natural disasters. goveent and non-goveent organizations, there e usually problems of unfair allocation and unsuitable vehicles for rescuing. Therefore, if there exist some mechanisms to integrate and allocate the information about rescuing resources when disasters occur, the rescuing process would be more efcient and effective. The emergency logistics planning has been studied for a long time and very benecial to human being in real life. All of the problems are somewhat different om each case that mapped into different environment, population distribution, transportation network, relief needs and geography situation. These factors mentioned before make it hard to compare with each other. From the defmition of San Diego State University, the logistics section of the emergency plan contains the service providing of food, facilities, human resource, and transportation. From the related denition above, this paper aims at providing the service at the se time. Those problems are usually mixed with many subproblems such as emergency rescue under the limited 72 hours for the experience of human physical ultimate, combining the VRP (Vehicle Routing Problem) [1]-[3] to dispatch emergency resource, limited capacity, the PDP (pickup-delivery problem) with different depots or destinations, time window constraints, that all make the problem even more complex. Based on the subproblems mentioned above, the research has several similar objectives such as saving time, lower economical cost, sending relie saving lives, etc. Also, there are many types of solutions such as dynamic integer linear progr m ing [4]-[10], Dijkstra algorithm [11]-[13], metaheuristic algorithms as ant colony and genetic algorithms [11], [13]-[17], immune intelligence [18], greedy method [19], game theo [20], zzy multi-objective program [4], [16], [21]-[23], etc. Aſter the happening of a disaster, many processes need to be prepared and organized. Before the step of logistics plning, some literature to nd the best station to restore relief that locate near by the disaster area [24] or predict the amount of demand [4], [22]; In [23] Chen and Tzeng to repair the broken trsportation network to send relief. While ISBN 978-89-5519-163-9 881 Feb. 1922, 2012 ICACT2012