J. Badura, A. Laskowski, M. Antczak, S. Wasik
Hyper-heuristic is a special program that aims to solve optimization problems independently from their search space. Instead, it uses a set of so-called low-level heuristics that are simple algorithms dedicated to solving a specific problem. I would like to use crowdsourcing to collect low-level heuristics. I designed a protocol that enables people to submit their low-level heuristics and implemented it on the Optil.io platform. The system was tested by collecting over a thousand low-level heuristics for a flowshop optimization problem. It is expected that those low-level heuristics form several clusters of similar approaches. I researched several ways of clustering algorithms so that I can choose a set of representative low-level heuristics that hyper-heuristic will use.
Keywords: Hyperheuristic, Crowdsourcing, Combinatorial Optimization
Scheduled
FA2 Heuristics 1
June 11, 2021 9:15 AM
2 - LV Kantorovich
