Abstract
A sensitivity analysis method which is an extension of the conjugate direction method developed by Davidon, Fletcher and Powell is presented. Sensitivity means the change of an objective function caused by the deviation of independent variables in an arbitrary direction from the optimum point. Furthermore, by this method we can know whether a local optimum point has been really attained or not after termination of successive iterations.
A numerical example of successful application to an optimization problem in chemical engineering is shown.