Utility function based hybrid algorithm for channel routing

This paper presents a two layer channel router with no doglegs, based on a hybridization of Stochastic Evolution and Tabu search methods. The problem-domain knowledge expressed in the form of utility functions is used to guide the exploration of the search space. Unlike previous search heuristic based routers, the use of utility functions in our router provides a powerful tool to determine the best moves that guarantee convergence in shorter times. The algorithm begins with an initial placement of nets, generated such that the nets are in relative conformity with the vertical constraint graph. Vertical and horizontal constraints are observed during the search process. The feasibility of the ideas is demonstrated using five benchmark problems. Optimal solutions are found in each case