Abstract
When solving large scale zero-one programming problems with computers, computation is usually stopped in a given computation time, even if the final solution has not been reached. Therefore, it will be very convenient if computation time can be estimated. However, as the number of solutions explicitly examined is not certain in generally known methods for zero-one programming, the estimation of computation time is difficult, while the Primal Effective Gradient Method makes it possible, because the number of the repitition of the procedure is roughly known. This paper is intended to present an estimatin method of computation time for solving zero-one programming problems with the primal Effective Gradient Method, and the comparison of estimated and actual computation times.
