Roy J. Clinton

Method of forces for direction finding in interactive multicriterion optimization applications

The main purpose of this paper is to derive, illustrate, and validate a method of direction finding for use in multicriterion interactive optimization applications. A secondary purpose is to consider electronic spreadsheet operation as an instance of interactive multicriterion optimization and to test the new method in a spreadsheet model for aggregate production planning. The method derived here amounts to virtual direct specification of the gradient direction, but does so by an appeal to the simple physical notion of forces applied to the criteria. It is demonstrated that the idea is easily grasped and also gives effective performance in the application tested.

A mass concept for optimal scaling in the method of abstract forces

In the method of abstract forces for the direction finding steps of interactive multicriterion optimization, it has been found necessary to scale user responses. Previously an ad hoc constant scaling method, which has nevertheless been shown effective, has been used. In this paper, a concept of abstract mass is introduced. It is found that the previous ad hoc scaling method may be considered an approximation to the mass concept approach. Also the latter approach provides a sound theoretical basis for the scaling problem. A method for optimal scaling is proposed.

Interactive Optimization: An Optimal Scaling and Step Sizing Approach for the Method of Abstract Forces

The method of abstract forces (MAF), suggested by Troutt and Hemming (1985), for interactive multiple criteria optimization was modified to handle heterogeneous criteria. The modified method of abstract forces (MMAF) was tested by comparing solutions derived with it against solutions found with the more traditional “what-if” method. Both methods aid the decision maker (DM) in solving problems that involve numerous alternatives and conflicting criteria. A unique aspect of the MMAF is its capability of reverse problem solving (RPS). In RPS the DM works backward to determine the inputs that would result in the desired solution. In constrast a what-if method (WIFM) tells the DM what the results are after the DM varies the inputs.

Toward optimal scaling in the method of abstract forces for interactive multiple criteria optimization

This paper presents a multiple criterion decision making model based on the Troutt-Hemming (1984) Method of Abstract Forces (MAF). The MAF is an interactive optimization method that allows the units of measurement of the various criteria to be different. The proposed method uses a scaling technique that adjusts for the diversity in impacts caused by the differences in the units of measurement of heterogeneous criteria. The method was tested by comparing solutions derived with it against solutions found with the more traditional what-if problem solving model. The MAF, with the scaling algorithm, was shown to be far superior to the what-if method in the application tested.