Accelerated local Monte Carlo estimate in calculation of laser radiation fields in tissue

Lappa A.V., Kamalov V.A., Potapov A.E., Shipitsin I.E.

Saratov Fall Meeting '98: Light Scattering Technologies for Mechanics, Biomedicine, and Material Science: Proceedings of SPIE; V.V. Tuchin, et al., Eds.; 1999. Vol. 3726. P. 151-156.

The Monte Carlo methods are widely used in biooptics to calculate different characteristics of the optical radiation fields. In the most cases the radiation characteristics including local ones (fluence, absorbed dose, etc.) are calculated with non-local Monte Carlo estimates, which give the values averaged over some region. These methods can lead to large systematical errors if the characteristics have fast space variation. At the same time there are local algorithms in the theory of the Monte Carlo methods. They are widely used in calculations of fields of ionizing radiation. In particular, the famous Kalos `local estimate' allows computations of radiation flux at a fixed point. This estimate solves the problem of the systematical error but because of infinite variance it has lower convergence and stability in comparison with non-local methods. Recently a method of acceleration of convergence of the Monte Carlo estimates possessing infinite variance was suggested. In the present work this accelerated Monte Carlo algorithm is applied to the Kalos estimate under typical conditions of biooptics (rat liver irradiated by 1.064 mkm laser). It is shown that the algorithm gives the gain in computational time up to 100 times. Obtained results allow us to recommend this algorithm for calculation of local characteristics of laser radiation fields in biological tissues.