IRIS Pricing engine is a framework for computing forward MtMs in a coherent way.
Pricing Models: specified by- the underlying risk, diffusion equations, calibration method, PCA factors –dedicated hook functions driving the reduction of the dimension so only relevant factors are kept and performance calculation can be increased -, and correlations with other risks. In addition, a Malliavin weight formula can be specified. IRIS provides a Libor Market Model and a Black-Sholes implementation, but any client-specific pricing models can be added through a pricing API.
Simulation: A scenario grid is generated according to pricing model diffusion. Grid characteristics –number of steps, time step- are user-defined.
Pricing Method: Pricing is implemented using an American and/or a brute force Monte-Carlo method (by coupling backward induction with Monte Carlo simulations). The method is specified per instrument type and is user-defined. Pricing fully supports OIS discounting: discount indexes are derived per asset from a mapping table based on trade type, customer and trade CSA data. Interpolation –using a Brownian bridge for example- between dates is specified through the API. For brute force Monte-Carlo, pricing functions are required. IRIS proposes a pre-defined set for vanilla instruments. This set is API extendible.