Abstract
Some simulations by Genetic Algorithm discussed in the previous paper were carried out in order to optimize the harvesting sequence and starting date of it taking into consideration the characteristics of cropping types. GA solution was regarded as the optimum one by comparison with models derived from empirical considerations and actual conditions of harvesting. GA was effective to solve these problems because of the rapid convergence of solution. The following results were clarified by the characteristics of the optimum solution. It is better to harvest in the order from a matured cane. Considerable economical loss occurs by inadequate harvesting sequence. Fitness value draws a curve with a peak against to the starting date, therefore there is an optimum solution.
