Reinforcement learning optimisation for graded metamaterial design using a physical-based constraint on the state representation and action space
Abstract The energy harvesting capability of a graded metamaterial is maximised via reinforcement learning (RL) under realistic excitations at the microscale. The metamaterial consists of a waveguide with a set of beam-like resonators of variable length, with piezoelectric patches, attached to it. T...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-12-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-023-48927-3 |
| _version_ | 1797398075239563264 |
|---|---|
| author | Luca Rosafalco Jacopo Maria De Ponti Luca Iorio Richard V. Craster Raffaele Ardito Alberto Corigliano |
| author_facet | Luca Rosafalco Jacopo Maria De Ponti Luca Iorio Richard V. Craster Raffaele Ardito Alberto Corigliano |
| author_sort | Luca Rosafalco |
| collection | DOAJ |
| description | Abstract The energy harvesting capability of a graded metamaterial is maximised via reinforcement learning (RL) under realistic excitations at the microscale. The metamaterial consists of a waveguide with a set of beam-like resonators of variable length, with piezoelectric patches, attached to it. The piezo-mechanical system is modelled through equivalent lumped parameters determined via a general impedance analysis. Realistic conditions are mimicked by considering either magnetic loading or random excitations, the latter scenario requiring the enhancement of the harvesting capability for a class of forcing terms with similar but different frequency content. The RL-based optimisation is empowered by using the physical understanding of wave propagation in a such local resonance system to constrain the state representation and the action space. The procedure outcomes are compared against grading rules optimised through genetic algorithms. While genetic algorithms are more effective in the deterministic setting featuring the application of magnetic loading, the proposed RL-based proves superior in the inherently stochastic setting of the random excitation scenario. |
| first_indexed | 2024-03-09T01:19:29Z |
| format | Article |
| id | doaj.art-5f93f553c04d4d2bb3ab48883e45a43f |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-03-09T01:19:29Z |
| publishDate | 2023-12-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-5f93f553c04d4d2bb3ab48883e45a43f2023-12-10T12:13:31ZengNature PortfolioScientific Reports2045-23222023-12-0113111110.1038/s41598-023-48927-3Reinforcement learning optimisation for graded metamaterial design using a physical-based constraint on the state representation and action spaceLuca Rosafalco0Jacopo Maria De Ponti1Luca Iorio2Richard V. Craster3Raffaele Ardito4Alberto Corigliano5Department of Civil and Environmental Engineering, Politecnico di MilanoDepartment of Civil and Environmental Engineering, Politecnico di MilanoDepartment of Civil and Environmental Engineering, Politecnico di MilanoDepartment of Mathematics, Imperial College LondonDepartment of Civil and Environmental Engineering, Politecnico di MilanoDepartment of Civil and Environmental Engineering, Politecnico di MilanoAbstract The energy harvesting capability of a graded metamaterial is maximised via reinforcement learning (RL) under realistic excitations at the microscale. The metamaterial consists of a waveguide with a set of beam-like resonators of variable length, with piezoelectric patches, attached to it. The piezo-mechanical system is modelled through equivalent lumped parameters determined via a general impedance analysis. Realistic conditions are mimicked by considering either magnetic loading or random excitations, the latter scenario requiring the enhancement of the harvesting capability for a class of forcing terms with similar but different frequency content. The RL-based optimisation is empowered by using the physical understanding of wave propagation in a such local resonance system to constrain the state representation and the action space. The procedure outcomes are compared against grading rules optimised through genetic algorithms. While genetic algorithms are more effective in the deterministic setting featuring the application of magnetic loading, the proposed RL-based proves superior in the inherently stochastic setting of the random excitation scenario.https://doi.org/10.1038/s41598-023-48927-3 |
| spellingShingle | Luca Rosafalco Jacopo Maria De Ponti Luca Iorio Richard V. Craster Raffaele Ardito Alberto Corigliano Reinforcement learning optimisation for graded metamaterial design using a physical-based constraint on the state representation and action space Scientific Reports |
| title | Reinforcement learning optimisation for graded metamaterial design using a physical-based constraint on the state representation and action space |
| title_full | Reinforcement learning optimisation for graded metamaterial design using a physical-based constraint on the state representation and action space |
| title_fullStr | Reinforcement learning optimisation for graded metamaterial design using a physical-based constraint on the state representation and action space |
| title_full_unstemmed | Reinforcement learning optimisation for graded metamaterial design using a physical-based constraint on the state representation and action space |
| title_short | Reinforcement learning optimisation for graded metamaterial design using a physical-based constraint on the state representation and action space |
| title_sort | reinforcement learning optimisation for graded metamaterial design using a physical based constraint on the state representation and action space |
| url | https://doi.org/10.1038/s41598-023-48927-3 |
| work_keys_str_mv | AT lucarosafalco reinforcementlearningoptimisationforgradedmetamaterialdesignusingaphysicalbasedconstraintonthestaterepresentationandactionspace AT jacopomariadeponti reinforcementlearningoptimisationforgradedmetamaterialdesignusingaphysicalbasedconstraintonthestaterepresentationandactionspace AT lucaiorio reinforcementlearningoptimisationforgradedmetamaterialdesignusingaphysicalbasedconstraintonthestaterepresentationandactionspace AT richardvcraster reinforcementlearningoptimisationforgradedmetamaterialdesignusingaphysicalbasedconstraintonthestaterepresentationandactionspace AT raffaeleardito reinforcementlearningoptimisationforgradedmetamaterialdesignusingaphysicalbasedconstraintonthestaterepresentationandactionspace AT albertocorigliano reinforcementlearningoptimisationforgradedmetamaterialdesignusingaphysicalbasedconstraintonthestaterepresentationandactionspace |