Adapting Black-Scholes to a non-Black-Scholes environment via genetic programming

The authors propose a new methodology that uses genetic programming to approximate the relationship between option price, the terms of the option contract, and properties of the underlying stock price. A crucial advantage of the genetic programming approach is that one can include the Black-Scholes formula in the parameter set, which allows one to search for an approximation better than currently known formulae. Using Monte Carlo simulations, they show that when data is generated using a jump-diffusion process, genetic programming approximates the true solution better than Black-Scholes. Other advantages to the approach are its low demand for data and its computational speed