A computational evaluation of optimal solution value estimation procedures

https://doi.org/10.1016/0305-0548(88)90059-7Get rights and content

Abstract

This research focuses on evaluating optimal solution value estimation procedures for large scale optimization problems. For many types of problems a heuristic technique must be used, which means that the optimal solution value is not known, and can only be estimated. We develop two modifications to previous estimators of the optimal solution value and present computational results for these estimators, for problems that are typically solved by means of a heuristic technique. This is accomplished by applying the estimators to five different problem types, and computing the estimation errors. Both modifications performed extremely well on problems for which good heuristics are available.

References (24)

  • D.G. Dannenbring

    Procedures for estimating optimal solution values for large combinatorial problems

    Mgmt Sci.

    (1977)
  • A.M. Mood et al.

    Introduction to the Theory of Statistics

    (1974)
  • D.G. Dannenbring

    The evaluation of heuristic solution procedures for large combinatorial problems

  • M.H. Quenouille

    Notes on bias in estimation

    Biometrika

    (1956)
  • D.S. Robson et al.

    Estimation of a truncation point

    Biometrika

    (1964)
  • H.L. Gray et al.

    The Generalized Jackknife Statistic

    (1972)
  • R.L. Nydick

    Evaluating optimal solution value estimation procedures for large scale problems

  • K.L. McRoberts

    A search model for evaluating combinatorially explosive problems

    Opns Res.

    (1971)
  • B. Golden

    A statistical approach to the TSP

    Networks

    (1977)
  • B. Golden

    Point estimation of a global optimum for large combinatorial problems

    Commun. Stat.

    (1978)
  • P. Cooke

    Statistical inference for bounds of random variables

    Biometrika

    (1979)
  • F.B. Talbot et al.

    A comparative evaluation of heuristic line balancing techniques

    Mgmt Sci.

    (1986)

    • Cited by (5)

    ∗

    Robert L. Nydick Jr is Assistant Professor of Management at the College of Commerce and Finance, Villanova University, Pennsylvania. He earned a B.S. in operations research from the Philadelphia College of Textiles and Science, an M.S. in operations research from the University of Pennsylvania, and a Ph.D. in applied statistics from Temple University. His current research interests are in the area of heuristic techniques, including estimating optimal solution values, and inventory lot-sizing under dynamic demand conditions.

    †

    Howard J. Weiss is Professor of Operations Research at Temple University, Philadelphia, Pennsylvania. He earned a B.S. in applied mathematics/computer science from Washington University and an M.S. arid Ph.D. in industrial engineering/management science from Northwestern University. His publications have appeared in several journals including Management Science, Operations Research, and Naval Research Logistics Quarterly. In addition, he has coauthored Introduction to Mathematical Programming (Elsevier North-Holland, Amsterdam).

    View full text
    Copyright © 1988 Published by Elsevier Ltd.