Experimental study of Ant Colony Optimization Algorithms (ACOs) in the Python Optimisation Framework (POF) Master thesis topic proposal Superviser: Cristina Vieira ([email protected]) Introduction ACO algorithms are constructive metaheuristics working with populations of ants. Ants build incrementally solutions guided by pheromone trails and, possibly, by a heuristic function [1]. The Python Optimisation Framework (POF) is an independent optimisation problem platform for metaheuristics [2]. POF is a research and teaching tool that provides a clear separation between problems and solvers. Main contributions of POF are improving the understanding of the solvers and problems themselves, and enable a fairer assessment and comparison of the performance of different search algorithms. Project description One of the contributions of POF is integrate constructive search with local search, allowing metaheuristics to manipulate complete and incomplete solutions and to use different search paradigms. Due to the impact that the ACO algorithms have been in recent years, the main goal of this thesis is to implement some variant of ACO algorithms given particular relevance to the efficient implementation of incomplete solution neighbourhood generation. 11 References 1. M. Dorigo, T. Stutzle, Ant Colony Optimization, Bradford Book, 2004. 2. Cristina C. Vieira, “Uma Plataforma para a Avaliação Experimental de Metaheurísticas”, Tese de doutoramento, UAlg, Dezembro 2009. 3. Ant Colony Optimization Website, visited Dezember 2012, http://iridia.ulb.ac.be/~mdorigo/ACO/ACO.html. 21