CAJ | 학술논문

通过数据分析，结合传统灰色GM（1，1）模型的特点，基于遗传算法与新陈代谢思想提出了改进的GM（1，1，λ）模型。结果表明：GM（1，1）模型对分散数据预测精度较低，其精度等级为四级以下，最大相对误差大于45％，预测值逐年上升，与实际情况不符。而改进的GM（1，1，λ）模型的精度等级为三级，最大相对误差为18．716％，更好地反映了城市用水量的变化趋势，与观测数据最为接近，预测精度较高。
통과수거분석，결합전통회색GM（1，1）모형적특점，기우유전산법여신진대사사상제출료개진적GM（1，1，λ）모형。결과표명：GM（1，1）모형대분산수거예측정도교저，기정도등급위사급이하，최대상대오차대우45％，예측치축년상승，여실제정황불부。이개진적GM（1，1，λ）모형적정도등급위삼급，최대상대오차위18．716％，경호지반영료성시용수량적변화추세，여관측수거최위접근，예측정도교고。
Through data analysis and combined with the characteristics of traditional gray GM (1,1)mod-el, the paper presented improved GM (1,1,λ) model based on genetic algorithm and metabolism .The results indicated that prediction accuracy of GM (1,1) model is lower for scattered data , its accuracy lev-el is under the forth and the maximum relative error is 45%.The predictive value is increased year by year,which is inconsistent with the actual situation .The prediction accuracy level of improved GM (1,1,λ) model is the third and the maximum relative error is 18 .716%,which better reflects the change trend of urban water demand and is most close to the observation data ,and the prediction accuracy is higher .