Pareto-optimal cycles for power, efficiency and fluctuations of quantum heat engines using reinforcement learning
Paolo Andrea Erdman, Alberto Rolandi, Paolo Abiuso, Martí Perarnau-Llobet, Frank Noé
Abstract
A framework to fully optimize driven quantum heat engines is presented that allows identification of Pareto-optimal cycles that trade-off power, efficiency, and power fluctuations. The framework is then applied to a quantum-dot-based heat engine using reinforcement learning and analytical methods based on the slow- and fast-driving dynamical regimes.
Topics & Concepts
Reinforcement learningPareto principleReinforcementHeat engineQuantumPower (physics)Pareto optimalComputer scienceMathematical optimizationArtificial intelligenceMathematicsMulti-objective optimizationPhysicsEngineeringThermodynamicsQuantum mechanicsStructural engineeringAdvanced Thermodynamics and Statistical MechanicsThermal Radiation and Cooling TechnologiesAdvanced Thermodynamic Systems and Engines