An improved simulated annealing algorithm for interactive multi-objective land resource spatial allocation

The primary work of formulating land use planning is to pursuit an optimized land use pattern to guide human activities for land utilization effectively. However, due to numerous spatial units and multiple land use types, as well as the spatial heterogeneity and incompatible objectives, land resource spatial optimization (LRSA) becomes a challenging issue. Currently, relevant research focuses on the exploitation of methods to enhance the efficiency of LRSA to meet the interactive demand of plan making. This paper designed a new simulated annealing (SA) algorithm for LRSA and provided an interactive platform to determine the ideal land use pattern in terms of stakeholder preferences on conflicted objectives. First, a general optimization model with three parts, including objectives, constraints and the multi-objective decision making technique, was presented. Second, for SA with crossed combinations of three cooling functions and four types of solution renewal jump steps, 12 SA sub-models were proposed, of which efficiency was compared. Lastly Jiangdu County in China was used as a case study. The following conclusions were reached: our excavation made SA efficiency for LRSA increased by about 35%, and the sub-model with logistic curve as cooling function and jump step was a gradually decreasing parameter was the most effective model; the proposed approach could obtain an ideal solution on any location of the frontier according to stakeholder preferences for conflicted objectives, thereby providing a useful interactive tool to reach an agreeable scheme transparently.