2020/05/14 by Majid Mazouchi, Yongliang Yang, Mazouchi, Majid +3 · 1 citation
Computer Science · Engineering · #Adaptive Dynamic Programming Control #Advanced Control Systems Optimization #Control Systems and Identification #FOS: Electrical engineering #FOS: Mathematics #Optimization and Control (math.OC) #Systems and Control (eess.SY) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2005.07118
openalex publication_date 2020/05/14 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
This paper presents an iterative data-driven algorithm for solving dynamic\nmulti-objective (MO) optimal control problems arising in control of nonlinear\ncontinuous-time systems. It is first shown that the Hamiltonian functional\ncorresponding to each objective can be leveraged to compare the performance of\nadmissible policies. Hamiltonian-inequalities are then used for which their\nsatisfaction guarantees satisfying the objectives' aspirations. An\naspiration-satisfying dynamic optimization framework is then presented to\noptimize the main objective while satisfying the aspiration of other\nobjectives. Relation to satisficing (good enough) decision-making framework is\nshown. A Sum-of-Square (SOS) based iterative algorithm is developed to solve\nthe formulated aspiration-satisfying MO optimization. To obviate the\nrequirement of complete knowledge of the system dynamics, a data-driven\nsatisficing reinforcement learning approach is proposed to solve the SOS\noptimization problem in real-time using only the information of the system\ntrajectories measured during a time interval without having full knowledge of\nthe system dynamics. Finally, two simulation examples are provided to show the\neffectiveness of the proposed algorithm.\n