Speaker
Description
The goal of this work is to develop an algorithm for identifying the singularity in divergent multidimensional integrals of the given type $1/(x-c)$ in situations where $c$ is unknown beforehand, making it impossible to use GSL VEGAS type algorithms directly. A method has been created to find the singularity's location in the integration domain with respect to the delta value. This is achieved by identifying numerous points in the integration domain that do not exhibit a significant difference with the singularity. A binary search is used to approach the singularity itself, which greatly reduces the number of operations that need to be done and, by extension, the amount of time that needs to be spent working. As a result, it becomes possible to calculate the integral using the Monte Carlo integration method (classical version or GSL VEGAS). In addition, the paper presents a comparison with the previous version of the algorithm, where the search was reduced to determining singularity regions of a given size.
