Multiobjective decisions analysis for engineering systems

https://doi.org/10.1016/0305-0548(80)90028-3Get rights and content

Abstract

Two decision analysis methodologies are developed and compared for the evaluation and selection of optimal engineering systems, the method of complete orders and the method of dichotomies. The problem is viewed as consisting of seeking to rank a discrete set of engineering alternatives in a decreasing order of attractiveness, as determined on the basis of their performance over a discrete set of criteria. The first methodology proceeds by successive dichotomies of the set of alternatives until the irreducible subset (core) of equally attractive alternatives is isolated, whereas the latter seeks complete (linear) orderings of the alternatives. Both methodologies are based on a probability impact-matrix that reflects the pair-wise comparison of the alternatives. The evaluation of this matrix makes use of the first and second statistical moments of each attribute, with respect to each one of the alternatives. The hypotheses are tested that each alternative consecutively is superior to all other alternatives. The alternatives are ranked according to the relative robustness of these hypotheses. The quality of the ranking is also assessed by the level of inductive entropy achieved. The method of complete orders proved in general to be superior to the method of dichotomies.

References (18)

  • R.A. Howard

    Decision Analysis in Systems Engineering

  • K.R. McCrimmon

    An overview of multiple objective decision making

  • T. Baumgartner et al.

    A systems model of conflict and change in planning systems with multi-level, multiple objective evaluation and decision making

  • R.L. Keeney et al.

    Decisions with Multiple Objectives

    (1976)
  • B. Roy

    How outranking relation helps multiple criteria decision making

  • P. Bertier et al.

    Analyse des Données Multidemsionelless

    (1975)
  • J.M. Blin

    The general concept of multidimensional consistency: some algebraic aspects of the aggregation problem

  • S. Watanabe

    Information-theoretical aspects of inductive and deductive inference

    IBM J. Res. Development

    (1960)
  • A. Papoulis

    Probability, Random Variables, and Stochastic Processes

There are more references available in the full text version of this article.

Cited by (0)

View full text