An Agent-Based framework for modeling and solving location problems

The high relevance of location problems in the operations research literature arises from their wide spectrum of real applications, including decision optimization in industrial management, logistics, and territorial planning. Most of these optimization problems fall into the class of NP-hard problems, motivating the search for heuristic and approximated algorithms. Currently, a great interest is being devoted to those optimization approaches yielding a concrete integration with spatial analysis instruments (such as Geographical Information Systems) that provide the user with an easy visualization of input data and optimization results. Agent-Based computing was recently proposed as an alternative to mathematical programming in order to solve problems whose domains are concurrently distributed, complex, and heterogeneous, also thanks to the availability of many commercial and open source codes including graphical interfaces for the elements of the problem. In this paper we propose a general Agent-Based framework for modeling various location problems. Together with its description, we present some computational results confirming the suitability and the effectiveness of the proposed approach.