Implementing and testing the tabu cycle and conditional probability methods

摘要

The purpose of this paper is to describe the implementation and testing of the tabu cycle method and two variants of the conditional probability method. These methods were originally described in Glover and Laguna [Tabu search. Boston: Kluwer Academic Publishers; 1997] but have been largely ignored in the tabu search literature. For the purpose of testing, we employ a single-attribute implementation of a short-term memory procedure for the solution of a single machine scheduling problem. Computational experiments that employ instances with up to 200 jobs reveal the usefulness of the tabu cycle and the conditional probability methods as viable alternatives for managing the short-term memory in a tabu search implementation.