Censored Exploration and the Dark Pool Problem
This work addresses a specific problem in finance with incremental improvements over prior asymptotic methods.
The paper tackles the Dark Pool Problem in quantitative finance by introducing a multi-venue exploration algorithm for censored data, proving polynomial-time convergence to near-optimal policies and demonstrating effectiveness with real trading data.
We introduce and analyze a natural algorithm for multi-venue exploration from censored data, which is motivated by the Dark Pool Problem of modern quantitative finance. We prove that our algorithm converges in polynomial time to a near-optimal allocation policy; prior results for similar problems in stochastic inventory control guaranteed only asymptotic convergence and examined variants in which each venue could be treated independently. Our analysis bears a strong resemblance to that of efficient exploration/ exploitation schemes in the reinforcement learning literature. We describe an extensive experimental evaluation of our algorithm on the Dark Pool Problem using real trading data.