A VLSI Multiplication-and-Add Scheme Based on Swarm Intelligence Approaches
The basic tasks in Digital Signal Processing systems can be expressed as summation of products. In this paper such operation is analyzed in terms of parallel distributed computation starting from an improvement of the Modified Booth Algorithm able to avoid useless sub-operation in the process of multiplication. We show how a such reformulation can take advantages from the cooperation between cells of a small colony statistically achieving shorter computation time. The interaction among cells, based on simple social rules typical of Swarm Intelligence systems, leads to a full exploitation of cells computational capabilities obtaining a more efficient usage of their computational resources in a so important task. In this paper a preliminary VLSI implementation and theoretical results are presented to show the feasibility of this approach.
KeywordsFinite State Machine Swarm Intelligence Partial Product Control Layer Elaboration Layer
Unable to display preview. Download preview PDF.
- 4.Kawamura, H., Yamamoto, M.: Multiple Ant Colonies Algorithm Based on Colony Level Interactions. IEICE Trans. Fundamentals E83-A 2, 371–379 (2000)Google Scholar
- 5.Kube, C.R.: Collective Robotics: From Local Perception to Global Action. Ph.D. Thesis, University of Alberta (1997)Google Scholar
- 7.Martinoli, A., Yamamoto, M., Mondada, F.: On the Modelling of Bio-Inspired Collective Experiments with Real Robots. In: Proc. of the Fourth European Conference on Artificial Life (ECAL 1997), Brighton, UK (1997)Google Scholar
- 8.Parhi, K.K.: VLSI Digital Signal Processing Systems - Design and Implementation, pp. 484–489. Wiley-Interscience, Hoboken (1999)Google Scholar
- 12.Taillard, E.: Ant Systems. Technical Report IDSIA-05-99, IDSIA, Lugano (1999)Google Scholar