An •-approximation approach based on tunneling method is proposed for finding a global optimal solution of a function of several variables. In this approach, after some local minimum has been found, one must obtain a new initial point from which a better local minimum is searched by appropriate gradient method. For that, a Newton-like method called restoration procedure is used. In order to dispel the loitering phenomenon and cycles in this procedure, the authors adjust the angle of the current searching direction with the last and with a given vector. Numerical results of several standard test problems are presented. [In Japanese.]
