Optimal control of quantum thermal machines using machine learning

Abstract

Identifying optimal thermodynamical processes has been the essence of thermodynamics since its inception. Here, we show that differentiable programming (DP), a machine learning (ML) tool, can be employed to optimize finite-time thermodynamical processes in a quantum thermal machine. We consider the paradigmatic quantum Otto engine with a time-dependent harmonic oscillator as its working fluid, and build upon shortcut-to-adiabaticity (STA) protocols. We formulate the STA driving protocol as a constrained optimization task and apply DP to find optimal driving profiles for an appropriate figure of merit. Our ML scheme discovers profiles for the compression and expansion strokes that are superior to previously-suggested protocols. Moreover, using our ML algorithm we show that a previously-employed, intuitive energetic cost of the STA driving suffers from a fundamental flaw, which we resolve with an alternative construction for the cost function. Our method and results demonstrate that ML is beneficial both for solving hard-constrained quantum control problems and for devising and assessing their theoretical groundwork.

5 pages + 5 figures, SM: 3 page + 1 figure. Comments are welcomed!

References in corpus (10)

• Single ion heat engine with maximum efficiency at maximum power
• Quantum thermal machines and batteries
• Universal Work Fluctuations during Shortcuts To Adiabaticity by Counterdiabatic Driving
• Transient energy excitation in shortcuts to adiabaticity for the time dependent harmonic oscillator
• Analog of a quantum heat engine using a single-spin qubit
• The Quantum Refrigerator: The quest for absolute zero
• Speeding-up a quantum refrigerator via counter-diabatic driving
• Deep Reinforcement Learning for Efficient Measurement of Quantum Devices
• Protocol Discovery for the Quantum Control of Majoranas by Differentiable Programming and Natural Evolution Strategies
• Exact Delta Kick Cooling, Time-Optimal Control of Scale-Invariant Dynamics, and Shortcuts to Adiabaticity Assisted by Kicks

Cited by in corpus (23)

• Quantum Engines and Refrigerators
• Counterdiabatic Optimised Local Driving
• Pareto-optimal cycles for power, efficiency and fluctuations of quantum heat engines using reinforcement learning
• Dynamical heat engines with non--Markovian reservoirs
• Quadratic Enhancement in the Reliability of Collective Quantum Engines
• Optimal Thermometers with Spin Networks
• Dissipation-induced collective advantage of a quantum thermal machine
• Optimal Control in Disordered Quantum Systems
• Counterdiabatic optimized driving in quantum phase sensitive models
• Optimal finite-time heat engines under constrained control
• Experimental graybox quantum system identification and control
• Bath engineering enhanced quantum critical engines
• Optimization of Asymmetric Quantum Otto Engine Cycles
• Exploring the Optimal Cycle for Quantum Heat Engine using Reinforcement Learning
• Multi-Axis Control of a Qubit in the Presence of Unknown Non-Markovian Quantum Noise
• Quantum optimal control in quantum technologies. Strategic report on current status, visions and goals for research in Europe
• Differentiable master equation solver for quantum device characterisation
• Parametrized multiqubit gate design for neutral-atom based quantum platforms
• Quantum compiling with a variational instruction set for accurate and fast quantum computing
• Preparing Quantum States by Measurement-feedback Control with Bayesian Optimization
• Machine learning delta-T noise for temperature bias estimation
• Stochastic Thermodynamics at the Quantum-Classical Boundary: A Self-Consistent Framework Based on Adiabatic-Response Theory
• Detecting Markovianity of Quantum Processes via Recurrent Neural Networks