Parallel computers and parallel algorithms have increasingly found their way into metaheuristics. Most parallel implementations of GRASP found in the literature consist in either partitioning the search space or the GRASP iterations and assigning each partition to a processor. GRASP is applied to each partition in parallel. These implementations can be categorized as multiple-walk independent-thread, with the communication among processors during GRASP iterations being limited to the detection of program termination and gathering the best solution found over all processors. Approaches for the parallelization of GRASP with path-relinking can be categorized as either multiple-walk independent-thread or multiple-walk cooperative-thread, with processors sharing and exchanging information about elite solutions visited during the GRASP iterations. This chapter is an introduction to parallel GRASP heuristics, covering multiple-walk independent-thread strategies, multiple-walk cooperative-thread strategies, and some applications of parallel GRASP and parallel GRASP with path-relinking.
CITATION STYLE
Resende, M. G. C., & Ribeiro, C. C. (2016). Parallel GRASP heuristics. In Optimization by GRASP (pp. 205–227). Springer New York. https://doi.org/10.1007/978-1-4939-6530-4_10
Mendeley helps you to discover research relevant for your work.