The Limitations of Frequency Analysis for Dendritic Cell Population Modelling
In previous work we derived a mathematical model which allows the frequency response of a dendritic cell to be predicted. The model has three, key limitations: the model assumes that the intermediate co stimulatory molecule signal is constant; it is only possible to make predictions for a single cell and the model only takes into account the signal processing element of the dendritic cell algorithm, with no attempt to explore the antigen presenting phase. In this paper we explore the original model and attempt to extend it to include the effects of multiple cells. It is found that the complex interactions between the cells creates a one to many relationship between the input frequency and the output frequency. This suggests that traditional frequency-based techniques alone are unlikely to yield an effective automated tuning mechanism.
KeywordsDendritic Cell Algorithm Frequency Analysis Tuning
Unable to display preview. Download preview PDF.
- 1.Greensmith, J., Aickelin, U., Cayzer, S.: Introducing dendritic cells as a novel immune inspired algorithm for anomaly detection. In: Jacob, C., Pilat, M.L., Bentley, P.J., Timmis, J.I. (eds.) ICARIS 2005. LNCS, vol. 3627. Springer, Heidelberg (2005)Google Scholar
- 3.Greensmith, J., Twycross, J., Aickelin, U.: Dendritic cells for anomaly detection. Congress on Evolutionary Computation (CEC) (2006)Google Scholar
- 4.Greensmith, J.: The Dendritic Cell Algorithm. PhD Thesis. The University of Nottingham (2007)Google Scholar
- 5.Ifeachor, E.C., Jervis, B.W.: Digital Signal Processing: A Practical Approach. Prentice-Hall, Englewood Cliffs (2001)Google Scholar