Approximation of Convex Envelope Using Reinforcement Learning
This addresses a mathematical optimization problem for researchers in computational geometry and machine learning, but appears incremental as it builds on an existing stochastic control formulation.
The paper tackled approximating the convex envelope of non-convex functions by developing a reinforcement learning scheme based on Q-learning for optimal stopping, showing promising results on standard test problems.
Oberman gave a stochastic control formulation of the problem of estimating the convex envelope of a non-convex function. Based on this, we develop a reinforcement learning scheme to approximate the convex envelope, using a variant of Q-learning for controlled optimal stopping. It shows very promising results on a standard library of test problems.