A diagnostic assessment of evolutionary algorithms for multi-objective surface water reservoir control

摘要: Abstract Globally, the pressures of expanding populations, climate change, and increased energy demands are motivating significant investments in re-operationalizing existing reservoirs or designing operating policies for new ones. These challenges require an understanding tradeoffs that emerge across complex suite multi-sector river basin systems. This study benchmarks our current capabilities to use Evolutionary Multi-Objective Direct Policy Search (EMODPS), a decision analytic framework which reservoirs’ candidate represented using parameterized global approximators (e.g., radial basis functions) then those functions optimized multi-objective evolutionary algorithms discover Pareto approximate policies. We contribute comprehensive diagnostic assessment modern MOEAs’ abilities support EMODPS Conowingo reservoir Lower Susquehanna River Basin, Pennsylvania, USA. Our results highlight can be very challenging some MOEAs epsilon dominance, time-continuation, auto-adaptive search helpful attaining high levels performance. The ϵ-MOEA, Borg MOEA, ϵ-NSGAII all yielded superior six-objective benchmarking test case. top show low sensitivity different MOEA parameterization choices algorithmic reliability consistent random trials. Overall, poses promising method discovering key management tradeoffs; however choice remains concern problems increasing complexity.