Informed multi-objective decision-making in environmental management using Pareto optimality.
Effective decision-making in environmental management requires the consideration of multiple objectives that may conflict. Common optimization methods use weights on the multiple objectives to aggregate them into a single value, neglecting valuable insight into the relationships among the objectives in the management problem. We present a multi-objective optimization procedure that approximates the non-dominated Pareto frontier without the use of weightings, allowing for visualization of the trade-offs among objectives. The non-dominated Pareto frontier is approximated by the simultaneous optimization of a vector objective function; two vector objective functions are defined as non-dominated if improvement with respect to one objective is at the detriment of another objective. We demonstrate the method with a case study for the optimum distribution of forest fuels treatments that reduce the impact of fire on a forest. The multiple objectives are to protect habitat of an endangered species, protect late successional forest reserves and minimize the total area treated. In the comparison of three optimization searches, the number of non-dominated solutions increases with the dimensions of the objective space, but with only two objectives the search is ineffective in minimizing fire impact in the different landscape types. Key challenges include the extensive computation time required to approximate the non-dominated set, and reducing the number of solutions that are analysed in detail. Synthesis and applications. The multi-objective optimization program presented can be adapted to other environmental management problems, and easily incorporates a wide range of quantifiable objectives. This tool provides decision-makers with a set of alternatives that estimates the full range of trade-offs among multiple objectives and provides a common ground from which dialogue can come to an informed compromise and decision in environmental management problems.