Directing against problems of using grey model GM(1

1) based on genetic algorithm in electric load prediction for a short term

namely it can't effectively reflect periodic variation of the load

and hasn't high accuracy of the load prediction

the model GM(1

1) has been extended as a model GM(1

1

λ)

and the solution of optimal value λ being found by the genetic algorithm

at the same time

the said model having been used in one power gride of Henan Province for practical prediction of load in 24h of the future. Result shows

the model (1

1

λ) based on genetic algorithm to have higher accuracy of load prediction

effectiveness of prediction being obvious.