scientific article; zbMATH DE number 3850824

From MaRDI portal
Publication:3320126





[For the entire collection see Zbl 0521.00029.] Let \(z^*\) be the optimal value of a combinatorial optimization problem. Any feasible solution z with probability \(p(z>z^*)<1-\alpha\), where \(\alpha\in(0,1)\), is called a statistical lower bound with confidence limit 1-\(\alpha\). The author describes how to use these bounds in the evaluation of heuristics, and in branch-and-bound procedures. He uses the Weibull distribution function to compute the bounds. Results obtained for the traveling salesman problem and the quadratic assignment problem are presented. A more detailed reference on the subject by the same author is On the use of confidence limits for global optimum in combinatorial optimization problems, Report No.83295-OR, Inst. für Oper. Res., Univ. of Bonn (1983).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3320126)