Optimal Execution via Multi-Objective Multi-Armed Bandits (Student Abstract)

When trying to liquidate a large quantity of particular stock, the price that stock is likely be affected by trades, thus leading reduced expected return if we were sell entire at once. This leads problem optimal execution, where aim split order into several smaller orders over course period time, optimally balance with market risk. can defined in terms difference equations. Here, show how reformulate this as multi-objective problem, which solve novel multi-armed bandit algorithm.