基于遗传算法的可持续土地利用动态规划

在建立和求解多目标可持续土地利用动态规划模型过程中，采用遗传算法并结合多目标模糊优选理论处理多目标土地利用结构优化问题。其详细步骤为：在分析区域特征基础上构建时间序列；通过现状分析设置土地利用类型决策变量，确定优化目标和约束条件，构建土地利用结构多目标优化模型；用灰色GM(1,1)预测模型对土地利用结构优化多目标模型参数进行预测；用遗传算法多目标优化方法对土地利用结构优化模型求解得pareto解集；最后运用多目标模糊优选动态规划方法对规划期内各时间段的pareto解集进行模糊动态优选，以最大优属度判定原则确定出最优决策序列。自贡市的实例表明，该方法有助于解决多目标协调问题，并通过提供多方案选择性，从而使规划更具灵活性，同时又便于充分吸纳公众参与，减少规划中的不确定性，增加规划的科学性、现实性和可操作性。

DYNAMIC PROGRAMMING METHOD FOR LAND SUSTAINABLE UTILIZATION BASED ON GENETIC ALGORITHM

A method of dynamic and multi-objective programming for land sustainable utilization is suggested based on genetic algorithm in combination with multi-objective fuzzy optimization.The logical approaches comprise of:(1) building time series based on careful analyzing of land utilization system,(2) setting up land use pattern variable and determining optimizing goal and the restriction condition from present situation,(3) applying grey GM(1,1) to predict parameters needed by the model,(4) studying pareto non-inferior desegregations by genetic algorithm,and(5) seeking optimum planning decision array from pareto desegregations by means of principle of maximum dominance.An instance of sustainable land use planning in Zigong region passing through time slicing 2010,2015 and extending into 2020 shows that the genetic algorithm integrated in multi-objective fuzzy optimum dynamic programming is conductive to solve the multi-objective coordination by providing with more choice schemes that in the meantime also helps to make much more flexibility and expedience to public participants in the process of planning.All procedures which produce outcomes of planning take consideration of characters of scientificalness,feasibility and maneuverability.