Ant Colony Based Algorithms for Dynamic Optimization Problems

The use of metaheuristic approaches to deal with dynamic optimization problems has been largely studied, being evolutionary techniques the more widely used and assessed techniques. Nevertheless, successful applications coming from other nature-inspired metaheuristics, e.g., ant algorithms, have also shown their applicability in dynamic optimization problems, but received a limited attention until now. Different from perturbative techniques, ant algorithms use a set of agents which evolve in an environment to construct one solution. They cooperate by means of asynchronous communications based on numerical information laid on an environment. This environment is often modeled by a graph which constitutes a formalism with a great expressiveness, specially well-suited for dynamic optimization problems. A solution could be a structure like a subgraph, a route, a short path, a spanning tree, or even a partition of vertices. In this chapter we present a general overview of the more relevant works regarding the application of ant colony based algorithms for dynamic optimization problems. We will also highlight the mechanisms used in different implementations found in the literature, and thus show the potential of this kind of algorithms for research in this area.