An interactive multiple objective linear programming method based on piecewise-linear additive value functions. (English) Zbl 0644.90086

The procedure proposed in this paper can be seen as an extension of the later version of the Zionts-Wallenius algorithm [cf. S. Zionts and J. Wallenius, Manage. Sci. 29, 519-529 (1983; Zbl 0519.90083)]. It uses the idea of approximating a utility function by a piecewise-linear additive form introduced in the UTA method by Jacquet-Lagrèze and Siskos [cf. E. Jacquet-Lagrèze and J. Siskos, Eur. J. Oper. Res. 10, 151-164 (1982; Zbl 0481.90078)]. Provided that the DM’s preferences can be represented by an additive value function with concave terms, the method, through a series of pairwise comparisons in the Zionts-Wallenius sense, allows the DM’s most preferred solution to be approximated, even when this is not a basic solution to the original problem.
It is worth noting that, independently, another procedure has been proposed, based on an interactive assessment of a piecewise-linear utility function for multiobjective linear programming [cf. E. Jacquet-Lagrèze, R. Meziani and the reviewer, ibid. 31, 350-357 (1987; Zbl 0636.90083)]. Instead of using the scheme of the Zionts-Wallenius algorithm, it uses a linear regression method for approximating the DM’s preferences by an additive utility function.
Reviewer: R.Słowinski


90C31 Sensitivity, stability, parametric optimization
91B16 Utility theory
90B50 Management decision making, including multiple objectives
90C05 Linear programming
Full Text: DOI