There are a huge amonut of problems with truly interest in science and engineering that can be formulated as an Optimization Problem. Some of these problems are solvable by Exact Algorithms. Branch and bound is a general strategy for finding optimal solutions to optimization problems. It consists of a systematic enumeration of all solutions, where a large number of them are discarded by using upper and lower bounds of their objective function value.
However, not all problems can be approached by an Exact Algorithm, at least in a reasonable running time. In order to solve these "dificult" problems metaheuristics can be used, obtaining high quality solutions in short running time. For the sake of brevity, only a subset of metaheuristics is presented, which represents a good sample of the different search paradigms proposed in the field, namely: the Greedy Randomized Adaptive Search Procedure (GRASP), Simulated Annealing, Tabu Search, Variable Neighborhood Search (VNS), Ant Colony Optimization, Evolutionary Algorithms, Scatter Search and Path Relinking.
The main research of the group focuses on the development of metaheuristic procedures for solving hard (combinatorial and continuous) optimization problems but we have also worked on exact methods for some selected problems.
Our research group supports a website (Optsicom Project) which contains a formal definition, the standard benchmark, and the best known results for all optimization problems where we are involved in.