I don’t know whether you said that a CEP application must necessarily have a model. It may have, or it may not. A rule-based approach (in its general acceptation) is not considered as a model. In the AI terminology, rules are considered as “shallow knowledge”, while models are considered as “deep knowledge”. Shallow knowledge expresses the people’s experience, links symptoms to causes directly, while deep knowledge establishes the links using a model, and the model can be interpreted. Shallow knowledge is very helpful in many cases, and as deep knowledge it also allows detecting situations. Of course, the cooperation of both is desirable to build more powerful systems. I did a rapid search, and below are 3 entries for reference:
E. Leonardi, F. Abel, D. Heckmann, E. Herder, J. Hidders, and G. Houben. Proceedings of 10th International Conference on Web Engineering (ICWE 2010), Vienna, Austria, July 5-9, 2010, volume 6189 of Lecture Notes in Computer Science, page 322-336. Springer, (2010)
D. Nguyen, D. Nguyen, and S. Pham. Proceedings of the 25th International Conference on Industrial Engineering and Other Applications of Applied Intelligent Systems, page 156-165. Springer-Verlag LNCS/LNAI, (2012)
D. Nguyen, D. Nguyen, and S. Pham. Proceedings of the 25th International Conference on Industrial Engineering and Other Applications of Applied Intelligent Systems, page 699-708. Springer-Verlag LNCS/LNAI, (2012)
M. Hearst. Proceedings of the 14th conference on Computational linguistics-Volume 2, page 539--545. Association for Computational Linguistics Morristown, NJ, USA, (1992)
W. Pedrycz, and M. Reformat. Proceedings of the Genetic and Evolutionary
Computation Conference (GECCO-2001), page 1389--1396. San Francisco, California, USA, Morgan Kaufmann, (7-11 July 2001)