遗传算法用于产生可供选择的城市规划方案

针对确定性的数学方法每次求解只能产生一个规划方案的问题 ,提出了产生可供选择的城市规划方案的遗传算法 .该遗传算法有着不同于其他文献的遗传编码、适应度测度和交叉的策略 .编码是用一个两维矩阵表示各类用地的空间分布 ,适应度是指各类用地单元的空间协调性 ,采用从上到下、从左到右顺序依次计算每个用地单元的空间协调性方法 ,交叉操作后进行个体检验使其满足约束条件 .利用该方法 ,在一个案例中 ,得到了多种规划方案 ,实验结果表明此方法是有效的、可行的

Using genetic algorithms to generate alternative schemes for urban planning

Deterministic mathematical method only produces one optimal solution in every computation in urban planning. This paper introduces a genetic algorithm(GA) designed to automate the process generating a set of alternative schemes for urban planning. Our GA has special encoding, mating, and calculating fitness's strategies not found in other genetic models. A two_dimensional matrix expresses the spatial distribution of urban land_use, and fitness value is defined as a sum of all cells' spatial harmony degree; and the computing process conforms to the order from top to bottom and from left to right in the matrix; and individuals should be examined after doing the cross_over operation in order to satisfy the constraints. By the application of our GA we can achieve many alternative sketch maps of land_use plan in a study area, and the performance of our GA shows it is powerful and efficient.