Stochastic multiple attribute evaluation of land use policies

Multiple attribute evaluation (MAE) of management actions is non-stochastic when the outcomes of those actions provide a fixed combination of attributes. Assuming the attributes of outcomes of management actions are non-stochastic and ranking those actions using MAE can lead to an inaccurate ranking when the attributes are stochastic. A stochastic MAE method is proposed for ranking management actions that have outcomes whose attributes are stochastic. The proposed method: (1) identifies possible actions for resolving a management problem; (2) selects and simulates the multiple attributes of the outcomes of possible management actions; (3) determines feasible management actions; (4) estimates attribute weights for different stakeholders; (5) uses stochastic dominance with respect to a function to rank feasible management actions based on the probability distributions of the attributes of outcomes; and (6) resolves conflicts in stakeholders' preferences for management actions. The proposed stochastic MAE method is demonstrated using a constructed example that uses simulated data to rank land use policies for stakeholders having different preferences for two interdependent attributes of the policy outcomes: buyer satisfaction with the density of residential housing and the percent of the area of interest with high realized habitat for multiple species. The method can be used to rank any set of management actions for which the stochastic attributes of outcomes can be characterized by probability distributions and the other information needed to apply the method is readily obtainable. Use of the method can be simplified by incorporating the various techniques and parameters into a spatial decision support tool.