Simulation Architecture Based on Distributive MDP for Inland Waterway Management

Climate change should affect inland waterways in a close future. The study of their resilience against climate change requires an optimal resource allocation. Indeed, it is possible to analyze the global change effects on inland waterways only if the water resources allocation is optimal. In addition, the events due to climate change are not deterministic. It is obvious that it is not possible to predict precisely their occurrence time, their magnitude and their duration. Hence, it is necessary to consider uncertainties in climate projections, and more precisely uncertainty bounds on these predictions. The objective of this paper is to propose a simulation architecture of inland waterways that couples simulation software of their dynamics and an optimal water resources allocation approach under uncertainties based on Markov Decision Process. The proposed simulation architecture and the designed tools are detailed and implemented by considering some expected climatic events on a part of the inland waterways in the north of France.