Identification of chaotic patterns in work processes

Abstract

During work, the performance of employees is not uniform, but fluctuates. In the case of algorithmic workflows, the goal is to keep this fluctuation as small as possible. The aim of the research was to reveal whether the fluctuations are random or organized into some kind of pattern. The research examined fluctuations in the performance of those performing algorithmic work based on their performance data over several months. During the investigation, we used the test of near returns, which is a topology-based procedure suitable for detecting small-dimensional chaotic behavior. The research showed that, in the case of the examined persons, performance fluctuations are not random, but individual patterns can be identified. With further targeted analysis, individual patterns provide an opportunity to reveal the hidden causes of chaotic behavior, which can reduce the fluctuation of workers' performance.