Revisiting Feller Diffusion: Derivation and Simulation
Ranjiva Munasinghe *
MIND Analytics and Management, 10/1 De Fonseka Place, Colombo 5, Sri Lanka.
Leslie Kanthan
University College London, Gower Street, London, WC1E 6BT, UK.
Pathum Kossinna
MIND Analytics and Management, 10/1 De Fonseka Place, Colombo 5, Sri Lanka.
*Author to whom correspondence should be addressed.
Abstract
We propose a simpler derivation of the probability density function of Feller Diffusion by using the Fourier Transform on the associated Fokker-Planck equation and then solving the resulting equation via the Method of Characteristics. We use the derived probability density to formulate an exact simulation algorithm whereby a sample path increment is drawn directly from the density. We then proceed to use the simulation to verify key statistical properties of the process such as the moments and the martingale property. The simulation is also used to confirm properties related to hitting time probabilities. We also mention potential applications of the simulation in the setting of quantitative finance.
Keywords: Feller diffusion, bessel squared process, simulation, absorbing random walks, survival probability.