A proposed alternative to phase-space recycling using the adaptive kernel density estimator method

Area of Science:

• Medical Physics
• Computational Physics
• Particle Accelerators

Background:

• Monte Carlo simulations are crucial for linear accelerator (linac) modeling.
• Accurate phase space (PS) data is essential for reliable linac simulations.
• Current methods may require enhancement for generating detailed PS distributions.

Purpose of the Study:

• To implement and evaluate the adaptive kernel density estimator (AKDE) for generating additional phase space (PS) variables in Monte Carlo linac simulations.
• To assess the accuracy and effectiveness of AKDE in enhancing simulated PS data.
• To improve the sampling of phase space points in accelerator simulations.

Main Methods:

• Implemented a nonparametric density estimation technique, the adaptive kernel density estimator (AKDE).
• Placed kernels with density-dependent window widths at simulated PS points.
• Reduced the original PS vector (x, y, u, v, E) to a rotationally invariant vector (r, theta, alpha, E) utilizing azimuthal symmetry.
• Sampled new PS vectors (r", theta", alpha", E") near original sampled vectors.
• Performed correlation analysis and compared distributions (fluence, energy, angular) between AKDE and original data.

Main Results:

• "In-air" particle fluence distributions agreed within 2% between AKDE samples and original distributions.
• Central axis energy and angular distributions showed agreement within 1.5% and 0.1%, respectively.
• Depth doses and dose profiles calculated using AKDE agreed within 2% and 2%/1 mm, respectively, compared to phase space recycling methods.
• AKDE demonstrated accuracy in reproducing key physical distributions.

Conclusions:

• The adaptive kernel density estimator (AKDE) is an effective method for generating additional phase space variables in Monte Carlo linac simulations.
• AKDE provides accurate and reliable enhancements to simulated phase space data.
• This technique shows promise for improving the precision and efficiency of accelerator simulations.