L'Hôpital's Filter for QSIM

Authors:
A. C. Cem Say
Affiliations:
Boğaziçi Univ., Istanbul, Turkey
Venue:
IEEE Transactions on Pattern Analysis and Machine Intelligence
Year:
1998

Citing 1
Cited 4

Improved Filtering for the QSIM Algorithm

IEEE Transactions on Pattern Analysis and Machine Intelligence

Sound and complete qualitative simulation is impossible

Artificial Intelligence
Qualitative simulation and related approaches for the analysis of dynamic systems

The Knowledge Engineering Review
Causes of ineradicable spurious predictions in qualitative simulation

Journal of Artificial Intelligence Research
Random generation of monotonic functions for Monte Carlo solution of qualitative differential equations

Automatica (Journal of IFAC)

Quantified Score

Hi-index	0.15

Visualization

Abstract

We have identified a source of spurious predictions inside the qualitative simulation algorithm QSIM. The algorithm fails to check for violations of l'Hôpital's rule, which causes the addition of inconsistent states to the behavior tree. Our proposed solution involves adding a new state filter to make the required controls and does not necessitate any additions or restrictions in the input set. We make use of extended corresponding value tuples spanning multiple constraints. The necessary modifications to the algorithm are explained and the technique is demonstrated on examples. Used in conjunction with other spurious behavior elimination methods, this approach would increase QSIM's ability to handle more complex systems.