An Eecient Multiprocessor Mapping Algorithmfor the Kohonen Feature Map and Its Derivative Models Contents 1 Introduction 1 2 Computational Parallelism in the Kohonen Model 1 3 Homogenous Multi-processor Architecture 2

This paper explores the potential for utilising specialised hardware for the implementation of the Kohonen model and its derivative models. The adaptation periods of these algorithms is potentially protracted and computationally expensive; especially for the Adaptive Kohonen model in real-world, on-line applications. This paper analyses these models and highlights inherent parallelism. This analysis forms the basis for examining the e cient implementation of these models using homogeneous, multi-processor computing architectures. The commonly used \network to processor" mapping algorithm is shown to be sub-optimal and an improved algorithm is proposed. The superiority of this new algorithm is demonstrated over the existing algorithms.