A Comparative Survey Of Software Computational Tools In The Field Of Optimal Control

Ožana, Štěpán; Docekal, Tomas; Nemcik, Jakub; Krupa, Filip; Możaryn, Jakub

doi:10.1109/pc52310.2021.9447510

Cited by 2 publications

(2 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the implementation of MPC the nonlinear optimization framework CasADi (23) is chosen due to the modular applicability, the large number of implemented optimizers and the very good benchmark results of related optimizer presented in (24) . For this study the direct multiple shooting method is used since it has better convergence than single shooting.…”

Section: Model Predictive Control Implementationmentioning

confidence: 99%

Comparison of Reinforcement Learning and Model Predictive Control for Automated Generation of Optimal Control for Dynamic Systems within a Design Space Exploration Framework

Hoffmann,

Gorelik,

Ivanov

2024

IJAE

View full text Add to dashboard Cite

This work provides a study of methods for the automated derivation of control strategies for over-actuated systems. For this purpose, Reinforcement Learning (RL) and Model Predictive Control (MPC) approximating the solution of the Optimal Control Problem (OCP) are compared using the example of an over-actuated vehicle model executing an ISO Double Lane Change (DLC). This exemplary driving maneuver is chosen due to its critical vehicle dynamics for the comparison of algorithms in terms of control performance and possible automation within a design space exploration framework. The algorithms show reasonable control results for the goal of this study, although there are differences in terms of driving stability. While Model Predictive Control first requires the optimization of the trajectory, which should then be optimally tracked, RL may combine both in one step. In addition, manual effort required to adapt the OCP problem to new design variants for solving it with RL and MPC is evaluated and assessed with respect to its automation. As a result of this study, an Actor-Critic Reinforcement Learning method is recommended for the automated derivation of control strategies in the context of a design space exploration.

show abstract

Section: Model Predictive Control Implementationmentioning

confidence: 99%

Comparison of Reinforcement Learning and Model Predictive Control for Automated Generation of Optimal Control for Dynamic Systems within a Design Space Exploration Framework

Hoffmann,

Gorelik,

Ivanov

2024

IJAE

View full text Add to dashboard Cite

show abstract

“…With such an approach, the solution obtained is not a set of time-equations, but a set of numerical vectors with the values of the states and controls describing the trajectory executed by the system in the time interval. The high computational cost related to the numerical approach is an inconvenience, furthermore, implementing such an approach is not trivial, despite this, the numerical approach has gained more and more space in research in relation to the analytical approach (RAO, 2010;OZANA et al, 2021).…”

Section: Solving the Optimal Control Problemmentioning

confidence: 99%

Direct Collocation Method for Solving Optimal Control Problems

Mosconi¹,

Luizete²,

Siqueira³

2023

Engenharias - Automação, Robótica, Metrologia E Energia: Estudos E Tendências - Volume 1

View full text Add to dashboard Cite

Objective: The purpose of this work was to present in details the development and use of the direct collocation method to solve optimal control problems. Methods: This was made in three parts: the first addresses how to formulate the optimal control problem, the second presents the direct collocation method addressing its construction, operation and comparison with other methods, and, finally, the third part introduces and example of an optimal control problem applied to a mechanical system (mass-spring-damper) and solved by the direct collocation method, making the algorithm used available to the reader. Results: The results obtained with the application of the optimal control to the mechanical system were the optimal states (position and velocity) and the optimal force needed to move the mass between three predefined points. Conclusion: With the work was possible to understand in detail the development and application of the direct collocation method to solve optimal control problem as well as to infer that it is a powerful and efficient tool when compared with other numerical methods.

show abstract

A Comparative Survey Of Software Computational Tools In The Field Of Optimal Control

Cited by 2 publications

References 11 publications

Comparison of Reinforcement Learning and Model Predictive Control for Automated Generation of Optimal Control for Dynamic Systems within a Design Space Exploration Framework

Comparison of Reinforcement Learning and Model Predictive Control for Automated Generation of Optimal Control for Dynamic Systems within a Design Space Exploration Framework

Direct Collocation Method for Solving Optimal Control Problems

Contact Info

Product

Resources

About