Maximum entropy estimates for risk-neutral probability measures with non-strictly-convex data (Q2247928)

The authors continue the work of \textit{J. Borwein} et al. [SIAM J. Optim. 14, No. 2, 464--478 (2003; Zbl 1050.91037)] on approximating the risk-neutral probability measure for the price of a financial asset at maturity via the maximum entropy principle. The problem is set up as a constrained optimization problem given a set of market observations and it is adapted for the case of a non-strictly convex data set. This type of data may accommodate better real noisy data. The optimization problem is then rigorously analysed via convex duality and constraint qualification on both bounded and unbounded price domains. The authors also show in detail how the arbitrage-free pricing can be obtained from risk-neutral measures in the case when the strict convexity condition is not fulfilled.