A Data-Driven Two-Stage Resource Allocation Model for Sewage Networks Against Failures

Sewage is a type of waste that is produced by domestic households, industrial and agricultural practices. The management of sewage networks is an important environmental issue, because the failure of network components results in sewage overflow to the ground surface, and thus creates offensive odors or even causes health hazards. In order to prevent such component failures, we address the optimal allocation of the limited resources by developing a data-driven two-stage non-linear programming model. In the first stage, the objective is to minimize the total allocated resources and the expected recovery costs of the network and environment, considering the random behaviors of component failures. In the second stage, the amount of leaked sewage caused by component failures is analyzed based on the hydraulic simulation model. The collected data is the historical failure data and detected network state for a fixed inspection interval. Driven by the data, the failure rate of network components is updated by the maximum likelihood estimation. Besides, genetic algorithm is proposed to analyze the two-stage model. In order to demonstrate the practical applicability of the proposed model, the case of a sewage network is conducted and some suggestions on the optimal resource allocations are finally provided. Furthermore, we discuss the impacts on the total costs caused by the data-driven updating of failure rate.