AntLib v1.0: A Generic C++ Framework for Ant Colony Optimization

摘要

This paper introduces AntLib, an Ant Colony Optimization (ACO) framework. C++ developed; it is a generic framework that can be applied with almost no adaptations to any combinatorial optimization problem. AntLib is a reusable object oriented framework, based on several well known design patterns [2]. AntLib is an extensible framework with a great amount of code ready to use, is efficient on execution leading to a high-speed performance, and it is very robust with loads of code checking. But the essential goal kept in mind in the AntLib design and development has been to provide a tool to the researcher that helps in the development phase of a combinatorial optimization system solved by ACO. AntLib offers a re-builder that reproduces a solution building process in order to achieve an easy debug and bug's repair tasks, a verifier to check whether the solution is properly built (to send it back to the rebuilder), and an analyst that studies the evolution of an algorithm, enables de parameters adjustment values and stagnation detection. AntLib is a generic framework, so almost no adaptation has to be done to solve different kinds of problems. Besides the object oriented paradigm, to achieve all the things mentioned before, AntLib makes use of templates, that can significantly reduce source code size and increases code flexibility without reducing safety type.