Peak traffic statistics and prediction for Markov process-based scheduling models

Yu Wu; Jianjun Yang

doi:10.1117/12.2684753

19 October 2023 Peak traffic statistics and prediction for Markov process-based scheduling models

Yu Wu, Jianjun Yang

Proceedings Volume 12709, Fourth International Conference on Artificial Intelligence and Electromechanical Automation (AIEA 2023); 1270961 (2023) https://doi.org/10.1117/12.2684753
Event: Fourth International Conference on Artificial Intelligence and Electromechanical Automation (AIEA 2023), 2023, Nanjing, China

Abstract

Many scheduling models have a finite number of states, mostly described using Markov processes. These models analyze the possible states of the system at different times. The scheduling algorithms for practical applications are combined with an engineering background and cannot be applied to other systems. The statistics and prediction of peaks also need to be combined with the needs of specific problems. In this paper, a peak state identification algorithm is designed based on Markov chains combined with machine learning methods. The algorithm can find the peak states from approximating the laws of the system itself in different cycles and can be used to predict the peak states in that cycle. We build a distance distribution to filter the suitable states and rank the different states by statistical analysis. We verify the accuracy of the algorithm by inputting random peak states, and the results show a high identification accuracy at different peak amounts. The algorithm can be applied to scheduling models with fixed nodes and with traffic demand and it has high accuracy in long runs.

Citation Download Citation

Yu Wu and Jianjun Yang "Peak traffic statistics and prediction for Markov process-based scheduling models", Proc. SPIE 12709, Fourth International Conference on Artificial Intelligence and Electromechanical Automation (AIEA 2023), 1270961 (19 October 2023); https://doi.org/10.1117/12.2684753

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available