Classical solutions to reaction-diffusion systems for hedging problems with interacting Itô and point processes (Q558663)

The authors are concerned with applications of reaction-diffusion systems to hedging and valuation problems in mathematical finance. Standard existence and uniqueness results for such systems may not apply because the coefficient functions of typical parametrizations in finance often are unbounded or do not satisfy linear growth constraints. Correspondingly, in Section 2 a first contribution of the present paper consists in deriving existence and uniqueness results for classical solutions of reaction-diffusion systems by showing existence of a unique fixed point of a suitable Feynman-Kac representation, provided that the underlying diffusion stays within a given domain and the coefficient functions satisfy local conditions. This first yields results under a global Lipschitz condition on the interaction term, and these are extended to a local Lipschitz condition by exploiting an additional monotonicity assumption on the interaction. In Section 3 a stochastic model with interacting Itô and point processes is introduced. It consists of a system of stochastic differential equations (describing the Markovian dynamics of a certain Itô process \(S\)) and a further finite-state process \(\eta\) (driven by the point processes). In Section 4, this framework is used as a rather flexible model for an incomplete financial market. Here, \(S\) describes the prices of tradable assets, while \(\eta\) represents further (not directly tradable) sources of financial risk.