;Due to the fact that Taiwan is located right on the pathway of typhoons from the western Pacific Ocean, typhoons frequently bring strong winds and heavy rain and it is easy to cause severe calamities. It might lead to loss and damage of human lives and properties. According to the statistics by the National Fire Agency, on average, there were 7.13 floods each year in the past 24 years. To prevent the levees from bursting and that people who live nearby will in danger, the Regulations on River Management stipulate that, in order to meet the needs in a flooding emergency, river management agencies must set up warehouses on appropriate sites along the river to store flood control and flooding emergency materials. “Flooding emergency” refers to measures taken to stop the situation from worsening as soon as flood control facilities have been damaged. In practice, the decision maker is used to deploy the flood control and flooding emergency materials based on his/her experience, which lakes optimal systematic analysis, Therefore, this research considers the stochastic demand occurring in actual situations, with the aim of optimizing the routing within the shortest period of time and minimizing costs, to construct flood control and flooding emergency materials deployment models. With these models, the decision maker can effectively deploy the flood control and flooding emergency materials at the warehouses.
In this research, the time-space network flow technique is used to construct the stochastic demand and deployment models. We further consider the average demand to construct the deterministic demand model. Both models are formulated as mixed integer multiple-commodity network flow problems, which are characterized as NP-hard. We utilize C++ computer language, coupled with the CPLEX mathematics programming solver, to solve the deterministic model. For the stochastic model, since their problem sizes are too huge to be directly solved by using mathematical programming software. Therefore, we developed a solution algorithm to efficiently solve the stochastic model. We also utilized EVPI and VSS to evaluate the performance of the stochastic model. Finally, we performed a case study using the data collected from a river management office. The test results show that the effectiveness achieved by applying the stochastic model is better than that by pragmatic decisions. The proposed model and solution algorithm could be useful for deploying the flood control and flooding emergency materials.