System design tools are often only available as blackboxes with complex nonlinear relationships between inputs and outputs . This paper presents CNMA (Constrained optimization withNeural networks, MILP solvers and Active Learning) It is conservative in the number of blackbox evaluations . It is resilient to the failure of blackboxes to compute outputs . CNMA improves the performance of stable, off-the-shelf implementations of Bayesian Optimization and Nelder Mead andRandom Search by 1%-87% for a given fixed time and function evaluation budget . It tries to sample only the partof the design space relevant to solving the design problem, leveraging the power of neural networks and MILPs, and a new learning-from-failure feedback loop . The paper also presents parallel CNMA that improves the efficiency and quality of solutions over the sequential version, and tries to steer it away from localoptima. It is evaluated for seven nonlinear design problems of 8 (2 problems), 10, 10, 15, 15 and 60 real-valued dimensions and one with 186 binary dimensions and it was evaluated for 7 nonlinear problems of eight nonlinear designs of 8 non linear design problems . It was shown that these implementations did not always return solutions. It’s shown that CNMA’s performance is evaluated to seven non

Author(s) : Sanjai Narain, Emily Mak, Dana Chee, Brendan Englot, Kishore Pochiraju, Niraj K. Jha, Karthik Narayan

Links : PDF - Abstract

Code :

https://github.com/oktantod/RoboND-DeepLearning-Project


Coursera

Keywords : design - nonlinear - problems - cnma - evaluated -

Leave a Reply

Your email address will not be published. Required fields are marked *