Google challenge: A hyperheuristic for the Machine Reassignment problem

In this work we present a hyperheuristic based method to solve Google Machine Reassignment problem (MRP). MRP was proposed at ROADEF/EURO challenge 2012 competition. It is a NP-complete problem. In the competition, this hard constrained optimization problem must be solved within 5 minutes. Our hyperheuristic approach uses a self-adaptive strategy according to the instance to solve, in order to quickly obtain quality solutions. The results show that self-adaptation is a good option for hyperheuristic approaches that require to be efficient in a restricted amount of time. Moreover, the results obtained using our hyperheuristic are competitive compared to those from the best algorithms of the competition.