When a large investor decides to liquidate his portfolio in finite time period, he can put market orders to do so. However large market orders may adversely impact prices. Which in turn may produce lower liquidation return . Dark pools are new kind of market, where not all the information is made public after the trade execution.
We have developed a dynamic strategy to place market orders as well as dark pool orders such that total expected liquidation return is maximized. The problem is formulated as Morkov Decision Process in finite horizon and solved using approximate dynamic programming technique.
Keywords : Portfolio Liquidation, Dark Pool, Dynamic Programming, Markov Decision Process
For details please read report.pdf.