A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2010; you can also visit the original URL.
The file type is
International Joint Conference on Neural Networks
We discuss methods for improving per3Pormance of the Hopfield-Tank quadratic minimization approach to TSP. A wide range of geometric (e.g., convex-hull based), topological and cutting-plane heuristics are investigated. We also investigate performance on non-Euclidean and non-metrizable TSP instances using a new concept of embedding dimension. Implications concerning the nature of the Hopjeld energy sulface are discussed. We conclude that the Hopfield-Tank formulation is not as robust as mightdoi:10.1109/ijcnn.1989.118627 fatcat:vyfeiwbhl5fi3jack3arwjproa