Remarks on the algorithm of sakawa for optimal control problems