交叉熵优化器：研究天体物理喷流中进动的新工具
Cross-entropy optimiser: a new tool to study precession in astrophysical jets
摘要：
在许多银河系和河外来源的喷流旋进的证据已在文献中报道。这些证据大部分是基于对喷流结运动学的研究，这取决于正确识别组件，以确定它们各自的自行和在天空平面上的位置角。与拟合程序有关的识别问题，以及采样时间不及时的观察结果，可能会影响及时跟踪组件，从而可能导致对数据的误解。为了克服这些局限性，我们引入了一个非常强大的统计工具来分析射流进动：连续多极值优化的交叉熵方法。交叉熵方法仅基于喷流分量的原始数据（距核心的赤经和赤经偏移）来搜索更好地代表数据的岁差模型参数。在这项工作中，我们提出了大量的测试，以验证这种技术，使用合成进动射流从一组给定的进动参数。为了恢复这些参数，我们将交叉熵方法应用于我们的进动模型，彻底改变与该方法相关的量。我们的结果表明，即使在最具挑战性的测试中，交叉熵方法也能够在1%的水平内找到正确的参数。即使对于非旋进射流，我们的优化方法也可以成功地指出旋进的不足。
英文摘要：
Evidence of jet precession in many galactic and extragalactic sources has been reported in the literature. Much of this evidence is based on studies of the kinematics of the jet knots, which depends on the correct identification of the components to determine their respective proper motions and position angles on the plane of the sky. Identification problems related to fitting procedures, as well as observations poorly sampled in time, may influence the follow up of the components in time, which consequently might contribute to a misinterpretation of the data. In order to deal with these limitations, we introduce a very powerful statistical tool to analyse jet precession: the cross-entropy method for continuous multi-extremal optimisation. Only based on the raw data of the jet components (right ascension and declination offsets from the core), the cross-entropy method searches for the precession model parameters that better represent the data. In this work we present a large number of tests to validate this technique, using synthetic precessing jets built from a given set of precession parameters. Aiming to recover these parameters, we applied the cross-entropy method to our precession model, varying exhaustively the quantities associated to the method. Our results have shown that even in the most challenging tests, the cross-entropy method was able to find the correct parameters within 1%-level. Even for a non-precessing jet, our optimization method could point out successfully the lack of precession.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
