Initial condition´s optimization on GM (1, 1)

A conventional GM(1,1) model solves grey differential equations through the first data of a raw sequence, and it has an effect on the precision of simulation and prediction. According to the principle of minimal information in grey system theory, this paper proposes an initial condition optimal approach of GM(1,1) model based on the minimal square sum on simulation error, The data relation between initial value and raw sequence is constructed through building a nonrestraint linear programming model on the basis of simulation error and initial value, and this paper also gives a calculation formula of initial value. Finally, an example demonstrates the validity and practicability of this approach.