基于遗传算法(GA)解甘蔗收割顺序最优化问题的研究

A genetic algorithm methodology for optimized harvesting sequence of sugarcane

本研究以已建立的栽培方法、收割时期预测甘蔗单产和含糖量的数学模型为基础,提出将寻求最佳收割顺序、收割期归结为最优化组合问题,并探讨了遗传算法(GA)的解题适用性,同时进行了几种基本设定的栽培方法推移、栽培方法构成项量、行列式特性的定性分析,为求最佳的甘蔗收割顺序、收割期、不同栽培方法的面积等提供解析方法。为不使栽培方法的构成比例发生变化,与通常GA算法采用交叉在不同个体间互相交替进行不同,开发了适用本研究的GA的自己交叉具体解析法。将收割第1年期作为过渡状态,第2年期为安定状态,将两期目的函数的合并作为一个适应度进行评价。

New mathematical models capable of predicting sugarcane sugar content based on planting method and harvest time were used to develop a methodology for optimizing harvesting sequence, and harvesting time of sugarcane. The applicability of GA was discussed too. The methodology based on Genetic Algorithm （GA） was used to qualitatively analyze the determinant variables of planting mode to obtain optimized combination harvest sequence, harvest time and area of different planting methods under several setted planting mode transitions. Self-crossing, as opposed to conventional GA crossing methods, was used to retain the structural ratio of planting modes. In this study, first year harvest was used for the transition state and that of the second year for the steady-state runs, while the combination of the two were formed the objective function.