Metaheuristics in Dynamic Environments
Metaheuristic techniques constitute a set of approximate algorithms which efficiently solve hard optimization problems. These algorithms have traditionally been applied to static optimization problems, i.e. problems which do not change while they are being solved. However many real world applications require to deal with changing environments: the estimation of the best route for a fleet of transport vehicles may depend on eventual breakdowns, weather broadcast and/or the state of traffic; software project planning could depend on changing requirements; market conditions in financial models are subject to change.
In order to solve this kind of problems using metaheuristic algorithms, it should be noted that the effectiveness of the algorithm can be improved if the algorithm retains some information about past states.
Last updates:
August 2011. The list of people working on metaheuristics and dynamic problems has been updated.
October 2009. A new VRP generator for dynamic instances is available for download here. This generator constructs VRP instances according to several customizable parameters, e.g. the number of customers, the size of the plane, the number of depots, the existence of time windows, etc. More important, it provides options to tune the dynamic problem according to many dynamic features: changing number of vehicles, changing lengths of routes, movement of the depot, existence of dynamic customers, etc.
University of Málaga
Department of Languages and Computation Sciences
NEO: Networking and Emerging Optimization