iNetwork A Framework for Network Optimisation

The use of heuristic search techniques for network optimisation problems has long been noted for its merits. It often requires both a significant understanding of both the heuristic methods themselves and the specific network problem domain. This precludes many network domain experts from effectively using heuristic search paradigms in building solutions. Solutions to network problems using heuristic methods generally involve significant development effort both to effectively drive the heuristic method and to model the domain requirements. Applying a different heuristic method to the same underlying problem often requires a major re-development of the original solution. The iNetwork framework presented in this paper provides a toolkit for the development of solutions to problems in network optimisation. iNetwork is an extension framework of the iOpt heuristic search toolkit and as such provides tools for modelling problems as invariant networks and the application of various heuristic search methods against those networks. iNetwork itself provides methods for specifying network problems in familiar network terms along with a broad base of algorithms that can be used in the formulation of solutions. Using a combination of adaptable network models, graph theory algorithms and heuristic based search algorithms, iNetwork provides a very flexible platform from which to develop solutions to network problems. The framework has initially been applied to the problem of Quality of Service routing. A short presentation of this work is given in the paper.