2010-2011

Summer term 2011

Title: Some non-Markovian control problems arising in order execution
Alexander Schied (University of Mannheim)
Wednesday, 11 May, 5:30pm, Room 139, Huxley Building

Abstract: We discuss several finite-fuel control problems that arise in connection with the optimal execution of orders under a risk criterion. These problems can be non-Markovian either through the price process, which may be a general semimartingale, or directly through the cost criterion, which may involve delay terms. In some cases, these problems can be solved explicitly via verification techniques. 


Title: An infinitely divisible distribution function useful for the cosmological many-body problem and (?) the financial many-body problem
Bill Saslaw (University of Virginia & Cambridge)
Wednesday, 18 May, 5:30pm, Room 139, Huxley Building


Title: Local volatility pricing models for long-dated derivatives in finance and insurance
Griselda Deelstra (Université Libre de Bruxelles) 
Wednesday, 25 May, 5:30pm, Room 139, Huxley Building 

Abstract: We study the local volatility function in the Foreign Exchange market where both domestic and foreign interest rates are stochastic. This model is suitable to price long-dated FX derivatives. We derive the local volatility function and obtain several results that can be used for the calibration of this local volatility on the FX option's market. We further concentrate upon Guaranteed Annuity Options (GAO), which give the right at the policyholder to convert his accumulated funds to a life annuity either at a rate based on the current market rates or at a fixed rate called the Guaranteed Annuity rate. We derive prices of GAO’s in the settings of a two-factor pricing model where the equity is locally governed by a geometric Brownian motion with a local volatility, while the interest rate follows a Hull-White one-factor Gaussian model.


Title: American option pricing via a probabilistic penalty method
Jan Palczewski (University of Warsaw and University of Leeds)
Wednesday, 1 June, 5:30pm, Clore LT (Level 2), Huxley Building

Abstract: In my talk I will present results concerning optimal stopping of functionals which exhibit discontinuity of the reward function at the boundary of an open (possibly unbounded) set. This type of functionals naturally arises when pricing American style options whose payoff changes abruptly when stock prices leave a given region. Of particular interest to me will be the continuity of the value function and the existence and form of optimal stopping times. These properties will be explored using the penalty method, an approach widely used in proving smoothness of solutions to QVI's. I will, however, apply penalization at the level of functionals, a method dating back to Robin (1978) and later mostly forgotten. Unjustly, as it works under weak assumptions (weakly Feller underlying processes) and gives rise to approximations of both the value function and an optimal stopping time. I will conclude with a discussion of numerical methods. This talk is partly based on a joint work with Lukasz Stettner.