Reliable two-stage location problem for mitigating flood impact on sustainable agricultural production and waste management

Sustainable agricultural production and waste management, i.e., stopping burning agricultural waste at fields and utilising the agricultural waste to produce biomass energy for improving the quality of environment, and enhancing the development of regional economy and society, have recently posed challenges due to impact of climate change (e.g., disruption risk by flooding on agricultural waste collection and transport network from fields to treatment facilities). In this paper, we develop a mixed-integer nonlinear programming model for the reliable agricultural waste collection, transport and treatment network design under impact of flooding that aims to minimise the impact of disruption risk on transport road network and locations of storages and facilities. The model can support rural planners to optimally locate storages (for collecting and transporting agricultural waste from fields to storages) and treatment facilities (for collecting and transporting agricultural waste from storages to treatment facilities). In the novel reliable two-stage location problem, the overall objective is to minimise total expected “disruption risk” cost. A solution technique based on “probability chains” is developed to transform the mixed-integer nonlinear programming model into a quadratic programming model that can be solved by advanced optimisation solvers. Based on the property of problem decomposition, we propose a two-stage solution method to seek quickly the optimal solutions of the reliable two-stage location problem. In addition, we develop an efficient math-heuristic algorithm for solving the large-sized instances within a reasonable computation time. The efficiency of the proposed solution methods is evaluated on the case studies in the Vietnam Mekong Delta.