Exploiting Synergies Between Exact and Heuristic Methods in Optimization: An Application to the Relay Placement Problem in Wireless Sensor Networks

Exact and heuristic methods for solving difficult optimization problems have been usually considered as two completely different approaches, with each of them being extensively studied by a different research community and used in different fields of application. In this work we propose a scheme for the integration and cooperation of an exact method, based on mixed-integer linear programming (MILP), and a bio-inspired metaheuristic, a genetic algorithm (GA), for the solution of a relay node placement problem in wireless sensor networks. The integration aspect relates to the use of a MILP solver inside GA’s operators. The cooperation aspect is implemented through a shared incumbent environment in which a MILP solver and the GA work in parallel and exchange relevant information in order to overcome their individual weaknesses and provide better solutions within a shorter time. Experimental results show a significant increase of performance when integration and cooperation take place, in comparison to the performance of the MILP and the GA when used independently from each other.