TY - GEN
T1 - Surrogates for Computationally Expensive and Failure-Prone Simulators
AU - Rooney, Sean
AU - Pitz, Emil
AU - Pochiraju, Kishore
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - Simulations used for navigating design spaces or finding optimal design points and Pareto fronts require accuracy and resolution to guide the designers towards effective decisions. High-fidelity or high-resolution simulators are computationally expensive. Due to inconsistent parameter settings and physically invalid outcomes, such simulators can also fail to return a solution during design automation or optimization loops. In CPS systems, failures can be expected when there is a hard execution-time deadline on the simulator for producing a solution. Time deadlines imposed by other considerations drive simulation failure behaviors in such cases.Current strategies for modeling simulator failure incorporate failures as an additional discrete parameter or entirely disregard the failed point. This paper describes a method for classifying candidate design points as predictable or failure-prone using a Bayes Classifier. Either a global surrogate or a mixture of the expert local surrogates will be identified in design spaces where the high-fidelity simulators yield a prediction. The developed technique is illustrated with two geometry assembly examples. The simulator fails if the composition does not lead to a valid 3D geometry or produces collisions in the assembly. This paper shows that global or MOE surrogates can be trained for both these cases with validation accuracy exceeding 90 %. The results also show that the best surrogate model can be a global model or a mixture of experts models and can vary by the approximated output parameter.
AB - Simulations used for navigating design spaces or finding optimal design points and Pareto fronts require accuracy and resolution to guide the designers towards effective decisions. High-fidelity or high-resolution simulators are computationally expensive. Due to inconsistent parameter settings and physically invalid outcomes, such simulators can also fail to return a solution during design automation or optimization loops. In CPS systems, failures can be expected when there is a hard execution-time deadline on the simulator for producing a solution. Time deadlines imposed by other considerations drive simulation failure behaviors in such cases.Current strategies for modeling simulator failure incorporate failures as an additional discrete parameter or entirely disregard the failed point. This paper describes a method for classifying candidate design points as predictable or failure-prone using a Bayes Classifier. Either a global surrogate or a mixture of the expert local surrogates will be identified in design spaces where the high-fidelity simulators yield a prediction. The developed technique is illustrated with two geometry assembly examples. The simulator fails if the composition does not lead to a valid 3D geometry or produces collisions in the assembly. This paper shows that global or MOE surrogates can be trained for both these cases with validation accuracy exceeding 90 %. The results also show that the best surrogate model can be a global model or a mixture of experts models and can vary by the approximated output parameter.
KW - Differentiable Surrogates
KW - Mixture of Experts
KW - Simulation Failure
UR - http://www.scopus.com/inward/record.url?scp=85134352712&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85134352712&partnerID=8YFLogxK
U2 - 10.1109/DESTION56136.2022.00007
DO - 10.1109/DESTION56136.2022.00007
M3 - Conference contribution
AN - SCOPUS:85134352712
T3 - Proceedings - 4th Workshop on Design Automation for CPS and IoT, DESTION 2022
SP - 1
EP - 6
BT - Proceedings - 4th Workshop on Design Automation for CPS and IoT, DESTION 2022
T2 - 4th Workshop on Design Automation for CPS and IoT, DESTION 2022
Y2 - 3 May 2022
ER -