Greed is good: Heuristic adaptations for resilience in renewable resource management

Abstract

Decision problems may be subject to objectives or constraints that make the formal model intractable. We propose an adaptive and pragmatic approach to address such nonstandard objectives or constraints, where these are first circumvented for feasibility, then accounted for through heuristics. One example is managing risk and resilience in a natural system facing uncertainty. Our procedure is exemplified in a predator‐prey fisheries system where a reference policy that maximizes expected profits implies a risk of prey stock collapse. The reference policy includes a no‐fishing section for the prey and harvest beyond myopic catch for the predator in parts of state space. We construct heuristic recovery plans, based on the reference policy, to reduce the risk of collapse by partly backing in the auxiliary objective. Under the heuristic policies, system resilience is enhanced with limited economic losses. Via Monte Carlo simulations, we calculate viability probabilities as measures of improved resilience and employ dynamic programming to assess value losses.