Ejection chain and filter-and-fan methods in combinatorial optimization

The design of effective neighborhood structures is fundamental to the performance of local search and metaheuristic algorithms for combinatorial optimization. Significant efforts have been made in the creation of larger and more powerful neighborhoods that are able to explore the solution space more extensively and effectively while keeping computation complexity within acceptable levels. The most important advances in this domain derive from dynamic and adaptive neighborhood constructions originating in ejection chain methods and a special form of a candidate list design that constitutes the core of the filter-and-fan method. The objective of this paper is to lay out the general framework of the ejection chain and filter-and-fan methods and present applications to a number of important combinatorial optimization problems. The features of the methods that make them effective in these applications is expected to provide insights into solving challenging problems in other settings.