Abstract
Large neighborhood search is a metaheuristic that has gained popularity in recent years. The heuristic repeatedly moves from solution to solution by
first partially destroying the solution and then repairing it. The best solution observed during this search is presented as the final solution. This
tutorial introduces the large neighborhood search metaheuristic and the variant adaptive large neighborhood search that dynamically tunes parameters
of the heuristic while it is running. Both heuristics belong to a broader class of heuristics that are searching a solution space using very large
neighborhoods.
The tutorial also present applications of the adaptive large neighborhood search, mostly related to vehicle routing problems for which the heuristic
has been extremely successful. We discuss how the heuristic can be parallelized and thereby take advantage of modern desktop computers that typically
contain several processing units and a software framework for implementing all of this is presented.
| Original language | English |
|---|---|
| Publication date | 2010 |
| Publication status | Published - 2010 |
| Event | MITACS / CORS 2010 Annual Conference - Edmonton, Canada Duration: 1 Jan 2010 → … |
Conference
| Conference | MITACS / CORS 2010 Annual Conference |
|---|---|
| City | Edmonton, Canada |
| Period | 01/01/2010 → … |