论文信息 - Measuring the Quality of Approximate Solutions to Zero-One Programming Problems

Measuring the Quality of Approximate Solutions to Zero-One Programming Problems

Abstract : The difficulties associated with measuring the are defined of an approximate (heuristic) solution by 'Percentage-Error' as is customary are pointed out. A set of properties that are to be expected from a proper measure of solution quality are defined and the family of functions which satisfy those conditions are investigated. In particular, it is shown that for any class of 0-1 programming problems appropriate functions do exist and that they are uniquely defined up to monotone transformations. Several examples of linear functions which are suitable for certain classes of problems are included.

Eitan Zemel

[1] Oscar H. Ibarra,et al. Fast Approximation Algorithms for the Knapsack and Sum of Subset Problems , 1975, JACM.

[2] George L. Nemhauser,et al. Note--On "Location of Bank Accounts to Optimize Float: An Analytic Study of Exact and Approximate Algorithms" , 1979 .

[3] Teofilo F. Gonzalez,et al. P-Complete Approximation Problems , 1976, J. ACM.

[4] Stephen A. Cook,et al. The complexity of theorem-proving procedures , 1971, STOC.

[5] Richard M. Karp,et al. On the Computational Complexity of Combinatorial Problems , 1975, Networks.

[6] David S. Johnson,et al. Approximation algorithms for combinatorial problems: an annotated bibliography , 1976 .

[7] David S. Johnson,et al. Approximation algorithms for combinatorial problems , 1973, STOC.