: To evaluate the effectiveness of an ant colony algorithm in implementing variance reduction techniques in the Monte Carlo computation of the photon beam quality correction factor for ionization chambers characterized with very small active volumes. The Monte Carlo code PENELOPE has been utilized to compute the photon beam quality correction factor for the Semiflex 3D 31021 ionization chamber, which has an active volume of . Various clinical photon beams generated with nominal potentials from 4 to MV have been considered, with a Co beam serving as the reference. The calculation involved determining the absorbed dose to both water and the sensitive volume of the ionization chamber. This information was used to derive the factors for the photon beams and the factor for the Co beam, whose ratio provides the factors. The algorithm has been initially validated by comparing with analog simulations where no variance reduction techniques are applied. The results have demonstrated an efficiency improvement ranging from a factor of 7 to 44. By incorporating the ant colony algorithm along with the variance reduction techniques, the determination of TPR values for various studied photon beams has been achieved. The calculated factors agree with previously published values. Two distinct protocols outlined in the TRS-398 have been taken into account and the results obtained for these protocols were compared to explore any differences between them. The ant colony algorithm facilitates the automatic application of variance reduction techniques, such as splitting and Russian roulette, without the need to delve into the geometric intricacies of the simulation. This automated approach results in increased efficiency, enabling simulations to be conducted within reasonable times while maintaining uncertainties at levels that ensure reliability.

中文翻译：

---

非常小体积电离室的光子束质量校正因子的蒙特卡罗计算[公式省略]：使用蚁群算法驱动的方差减少技术


：评估蚁群算法在蒙特卡罗计算中对具有非常小的活动体积的电离室的光子束质量校正因子实施方差减少技术的有效性。蒙特卡罗代码 PENELOPE 已用于计算 Semiflex 3D 31021 电离室的光子束质量校正因子，该电离室的有效体积为 .已考虑了以 4 至 MV 标称电势生成的各种临床光子束，并以 Co 束作为参考。计算涉及确定水的吸收剂量和电离室的敏感体积。该信息用于推导光子束的因子和钴束的因子，其比率提供因子。通过与未应用方差减少技术的模拟模拟进行比较，该算法已得到初步验证。结果表明，效率提高了 7 至 44 倍。通过将蚁群算法与方差减少技术相结合，可以确定各种研究光子束的 TPR 值。计算出的因子与之前公布的值一致。已考虑 TRS-398 中概述的两个不同协议，并对这些协议获得的结果进行比较，以探索它们之间的差异。蚁群算法有助于自动应用方差减少技术，例如分裂和俄罗斯轮盘赌，而无需深入研究模拟的复杂几何结构。 这种自动化方法可以提高效率，使模拟能够在合理的时间内进行，同时将不确定性保持在确保可靠性的水平。