Optimal Control using Evolutionary Algorithms through Neural network based TRANSFORMation

Miriyala, Srinivas S and Mitra, Kishalay (2020) Optimal Control using Evolutionary Algorithms through Neural network based TRANSFORMation. In: 2020 IEEE Symposium Series on Computational Intelligence, SSCI 2020, 1 December 2020 - 4 December 2020.

Full text not available from this repository. (Request a copy)


Conventional direct methods of solving Optimal Control (OC) problems lead to large scale optimization formulations, making the classical optimization solvers more preferable over evolutionary optimization algorithms, for solving the single and multiple objective formulations in OC. On the other hand, population based evolutionary optimization solvers have the ability to identify the global basin efficiently. Therefore, in this paper, a novel method termed as TRANSFORM Artificial Neural Network (ANN) assisted reformulation of OC, has been proposed, which transforms the large scale optimization problem into weight training exercise of auto-tuned ANNs that in turn reduces the scale of optimization by several folds. Through this reformulation, the implementation of evolutionary optimization algorithm is enabled for solving both single and multiple objective OC formulations. Three different benchmark case studies are considered from literature to test the efficiency of proposed algorithm -(a) control of a batch reactor for maximizing the yield of penicillin production, (b) optimal drug scheduling for maximizing the success rates in chemotherapy for cancer treatment, and (c) multi-objective control of plug flow reactor with energy and conversion trade-off. Results indicated an average 50-fold reduction in OC problem size due to ANN reformulation.

[error in script]
IITH Creators:
IITH CreatorsORCiD
Miriyala, Srinivas S.UNSPECIFIED
Mitra, Kishalayhttp://orcid.org/0000-0001-5660-6878
Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Benchmark case studies; Classical optimization; Control of a batch reactors; Evolutionary optimization algorithm; Evolutionary optimizations; Large-scale optimization; Multi-objective control; Penicillin production
Subjects: Chemical Engineering
Divisions: Department of Chemical Engineering
Depositing User: . LibTrainee 2021
Date Deposited: 28 Jun 2021 09:48
Last Modified: 28 Jun 2021 09:48
URI: http://raiith.iith.ac.in/id/eprint/8032
Publisher URL: http://doi.org/10.1109/SSCI47803.2020.9308475
Related URLs:

Actions (login required)

View Item View Item
Statistics for RAIITH ePrint 8032 Statistics for this ePrint Item