Pricing maximum-minimum bidirectional options in trinomial CEV model

Maximum-minimum bidirectional options are a kind of exotic path dependent options. In the constant elasticity of variance (CEV) model, a combining trinomial tree was structured to approximate the non-constant volatility that is a function of the underlying asset. On this basis, a simple and efficient recursive algorithm was developed to compute the risk-neutral probability of each different node for the underlying asset reaching a maximum or minimum price and the total number of maxima (minima) in the trinomial tree. With help of it, the computational problems can be effectively solved arising from the inherent complexities of different types of maximum-minimum bidirectional options when the underlying asset evolves as the trinomial CEV model. Numerical results demonstrate the validity and the convergence of the approach mentioned above for the different parameter values set in the trinomial CEV model.