Keywords: homeland security, emergence, patterns, terrorism, co-evolution, evolutionary computation, water distribution system.

The ultimate objective of the reported research is to develop proactive infrastructure security, which will include a better and more complete understanding of a given infrastructure system than a terrorist attack planner might have. The use of Information Technology can counterbalance the danger of unknown and unexpected terrorist threats. This paper focuses on the security improvements to an already existing infrastructure system, and identifies its vulnerabilities and available counter-measures. The proposed approach is based on the use of co-evolutionary algorithms, a concept from evolutionary computation, for the generation of both terrorist and security scenarios [2,3,4,7,11,12,13,14].

Water distribution systems are particularly vulnerable to terrorist attacks because of their ubiquity [5,8,9,10]. A hypothetical network for a small town has been modelled with EPANet [6], which is a public domain, water distribution system-modelling program developed by the U.S. Environmental Protection Agency's Water Supply and Water Resources Division. It was used as an evaluator in the broader framework of the evolutionary tool called SecurityMax/Water [1], developed in the Information Technology and Engineering School at George Mason University. The effects of contaminating and securing selected hydrants in the system have been studied.

A number of experiments, run with various evolutionary and co-evolutionary settings, are reported in the paper. The results of these experiments reveal emerging security patterns regarding the locations of "dangerous" fire hydrants, unique for a given water distribution system. Also, the interaction between terrorist and security scenarios is clearly visible. In the reported experiments, the terrorist scenarios were able to "avoid" security scenarios by employing only local modifications, however the security scenarios were able to follow and to some degree counteract the terrorist scenarios. Moreover, they were able to find dangerous regions in the network. Although the patterns of security scenarios were not entirely surprising, we can easily imagine a situation when the security scenarios use a totally different search space than the terrorist scenarios. In this context, the methodological results of our approach are more universal than the example presented in the paper. Understanding how different characteristics of a given problem influence the dynamics of co-evolution, as well as modelling situations in which terrorists modify their actions in response to possible security scenarios, are considered as further research directions.

1

T. Arciszewski, K.A. De Jong, "Evolutionary Computation in Civil Engineering: Research Frontiers" in B.H.V. Topping (ed), "Civil and Structural Engineering Computing: 2001", Saxe-Coburg Publications, Stirling, UK, 161-179, 2001. doi:10.4203/csets.5.7

2

T. Bäck, D.B. Fogel, Z. Michalewicz (eds), "Evolutionary Computation", Institute of Physics Publishing. Bristol, Philadelphia, 2000.

3

T. Bäck, D.B. Fogel, Z. Michalewicz (eds), "Handbook of Evolutionary Computation", IOP/Oxford Univ. Press. Amsterdam, Netherlands, 1997.

4

P. Bentley, "Evolutionary Design by Computers", Morgan Kaufmann, 1999.

5

G. Blumh, G. Nichols, E. Hawkins, M. Wadda, "Water Security Planning in Metropolitan Washington Area", COG Institute of Regional Excellence, project report, 1999.

6

U.S. Environmental Protection Agency EPANET 2000 webpage, www.epa.gov/ORD/NRMRL/wswrd/epanet.html (accessed February 13, 2005).

7

D.E. Goldberg, "Genetic Algorithms in Search", Optimization & Machine Learning, Addison-Wesley, Massachusetts, 1989.

8

Haestad Methods, Inc., Water Security Summit Proceedings, Haestad Methods, Inc., Connecticut, 2002.

9

L.W. Mays, "Water Distribution System Handbook", McGraw-Hill, New York, 2000.

10

L.W. Mays, "Water Supply System Security", McGraw-Hill, New York, 2004.

11

M. Mitchell, "An Introduction to Genetic Algorithms (Complex Adaptive Systems)", MIT Press, 1998.

12

I.C. Parmee, "Exploring the Design Potential of Evolutionary Search, Exploration and Optimisation", in "Evolutionary Design by Computers", P. Bentley, ed., Academic Press Ltd., London, 1999.

13

C.D. Rosin, "Coevolutionary Search Among Adversaries", Phd Thesis, University of California, San Diego, 1997

14

W.M. Spears, K.A. De Jong, T. Bäck, D.B. Fogel, H. De Garris, "An Overview of Evolutionary Computation", Proceedings of European Conference on Machine Learning, 442-459, 1993.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £80 +P&P)