Columbia & NYU Financial Engineering Colloquium: Jin Ma & David Yao

Registration is required. 

Jin Ma

University of Southern California

Title

Convergence Analysis of Policy Iteration Algorithms for Entropy-Regularized Stochastic Control Problems

Abstract

In this talk, I present a new scheme for the convergence analysis of the Policy Iteration Algorithm (PIA) for general continuous-time entropy-regularized stochastic control problems. The scheme builds on the probabilistic representation formulae for solutions as well as their derivatives of the family of iterative PDEs involved in the PIA, but without employing heavy PDE machinery. Such a scheme is particularly effective for the infinite horizon problems with large discounting factors, and we shall argue that, with some natural variations, it can be extended to the infinite horizon with a general discounting factor, the finite horizon case, and even some special cases when diffusion coefficient contains the control. Finally, we show that, in the infinite horizon model with a large discount factor and in the finite horizon model, the scheme can actually produce the exponential rate of convergence without tear.

This talk is based on the joint work with Gaozhan Wang and Jianfeng Zhang.

David Yao

Columbia University

Title

Emission Trading System and Risk Hedging

Abstract

Carbon emission has in recent years become an important issue in supply chains, specifically in the form of a cost on the input/supply side, like raw materials. Thus, when making production (or capacity) decisions, a firm has to take into account not only the risk involved in demand, but also the risk associated with carbon price, which fluctuates on the emission trading system (ETS).  We develop a hedging strategy by dynamically taking position on the ETS, and jointly optimize this strategy with the production-quantity decision. The model also demonstrates the positive impact of hedging on clean-tech adoption. (Joint work with Liao WANG, University of Hong Kong.)