Abstract
Mean-field models are an established method to analyze large stochastic systems with N interacting objects by means of simple deterministic equations that are asymptotically correct when N tends to infinity. For finite N, mean-field equations provide an approximation whose accuracy is model- and parameter-dependent. Recent research has focused on refining the approximation by computing suitable quantities associated with expansions of order $1/N$ and $1/N^2$ to the mean-field equation. In this paper we present a new method for refining mean-field approximations. It couples the master equation governing the evolution of the probability distribution of a truncation of the original state space with a mean-field approximation of a time-inhomogeneous population process that dynamically shifts the truncation across the whole state space. We provide a result of asymptotic correctness in the limit when the truncation covers the state space; for finite truncations, the equations give a correction of the mean-field approximation. We apply our method to examples from the literature to show that, even with modest truncations, it is effective in models that cannot be refined using existing techniques due to non-differentiable drifts, and that it can outperform the state of the art in challenging models that cause instability due orbit cycles in their mean-field equations.
- Michel Benaim and Jean-Yves Le Boudec. 2008. A class of mean field interaction models for computer and communication systems. Performance evaluation 65, 11--12 (2008), 823--838.Google Scholar
- Gunter Bolch, Stefan Greiner, Hermann de Meer, and Kishor Trivedi. 2005. Queueing networks and Markov chains: modeling and performance evaluation with computer science applications. Wiley.Google Scholar
- Luca Bortolussi, Jane Hillston, Diego Latella, and Mieke Massink. 2013. Continuous approximation of collective system behaviour: A tutorial. Performance Evaluation 70, 5 (2013), 317--349.Google Scholar
Digital Library
- Maury Bramson, Yi Lu, and Balaji Prabhakar. 2010. Randomized Load Balancing with General Service Time Distributions. In Proceedings of the ACM SIGMETRICS International Conference on Measurement and Modeling of Computer Systems. 275--286.Google Scholar
Digital Library
- Luca Cardelli, Mirco Tribastone, Max Tschaikowski, and Andrea Vandin. 2017. ERODE: A tool for the evaluation and reduction of ordinary differential equations. In International Conference on Tools and Algorithms for the Construction and Analysis of Systems. Springer, 310--328.Google Scholar
Digital Library
- Fabio Cecchi, Sem C Borst, and JSH van Leeuwaardena. 2015. Mean-field analysis of ultra-dense csma networks. ACM SIGMETRICS Performance Evaluation Review 43, 2 (2015), 13--15.Google Scholar
Digital Library
- Augustin Chaintreau, Jean-Yves Le Boudec, and Nikodin Ristanovic. 2009. The age of gossip: spatial mean field regime. 37, 1 (2009), 109--120.Google Scholar
- R.W.R. Darling and J.R. Norris. 2008. Differential equation approximations for Markov chains. Probab. Surveys 5 (2008), 37--79. https://doi.org/10.1214/07-PS121Google Scholar
Cross Ref
- Khanh N Dinh and Roger B Sidje. 2016. Understanding the finite state projection and related methods for solving the chemical master equation. Physical biology 13, 3 (2016), 035003.Google Scholar
- Stewart N Ethier and Thomas G Kurtz. 2009. Markov processes: characterization and convergence. Vol. 282. John Wiley & Sons.Google Scholar
- Nicolas Gast. 2017. Expected values estimated via mean-field approximation are 1/??-accurate. Proceedings of the ACM on Measurement and Analysis of Computing Systems 1, 1 (2017), 1--26.Google Scholar
Digital Library
- Nicolas Gast, Luca Bortolussi, and Mirco Tribastone. 2019. Size expansions of mean field approximation: Transient and steady-state analysis. Performance Evaluation 129 (2019), 60--80.Google Scholar
Digital Library
- Nicolas Gast and Gaujal Bruno. 2010. A mean field model of work stealing in large-scale systems. ACM SIGMETRICS Performance Evaluation Review 38, 1 (2010), 13--24.Google Scholar
Digital Library
- Nicolas Gast and Benny Van Houdt. 2015. Transient and steady-state regime of a family of list-based cache replacement algorithms. In Proceedings of the 2015 ACM SIGMETRICS International Conference on Measurement and Modeling of Computer Systems. 123--136.Google Scholar
Digital Library
- Nicolas Gast and Benny Van Houdt. 2017. A refined mean field approximation. Proceedings of the ACM on Measurement and Analysis of Computing Systems 1, 2 (2017), 1--28.Google Scholar
Digital Library
- Daniel T Gillespie. 2007. Stochastic simulation of chemical kinetics. Annu. Rev. Phys. Chem. 58 (2007), 35--55.Google Scholar
Cross Ref
- Andrew G Hart and Richard L Tweedie. 2012. Convergence of invariant measures of truncation approximations to Markov processes. (2012).Google Scholar
- MHR Khouzani, Saswati Sarkar, and Eitan Altman. 2012. Maximum damage malware attack in mobile wireless networks. IEEE/ACM Transactions on Networking 20, 5 (2012), 1347--1360.Google Scholar
Digital Library
- Juan Kuntz, Philipp Thomas, Guy-Bart Stan, and Mauricio Barahona. 2019. Stationary distributions of continuous-time Markov chains: a review of theory and truncation-based approximations. arXiv preprint arXiv:1909.05794 (2019).Google Scholar
- Thomas G Kurtz. 1978. Strong approximation theorems for density dependent Markov chains. Stochastic Processes and their Applications 6, 3 (1978), 223--240.Google Scholar
- Xin Liu, Lei Ying, et al. 2020. Beyond Scaling: Calculable Error Bounds of the Power-of-Two-Choices Mean-Field Model in Heavy-Traffic. arXiv preprint arXiv:2012.06613 (2020).Google Scholar
- Yuanyuan Liu and Wendi Li. 2018. Error bounds for augmented truncation approximations of Markov chains via the perturbation method. Advances in Applied Probability 50, 2 (2018), 645--669.Google Scholar
Cross Ref
- Yuanyuan Liu, Wendi Li, and Hiroyuki Masuyama. 2018. Error bounds for augmented truncation approximations of continuous-time Markov chains. Operations Research Letters 46, 4 (2018), 409--413.Google Scholar
Digital Library
- Laurent Massoulié and Milan Vojnovic. 2005. Coupon replication systems. In Proceedings of the 2005 ACM SIGMETRICS International Conference on Measurement and modeling of computer systems. 2--13.Google Scholar
Digital Library
- Hiroyuki Masuyama. 2017. Error bounds for last-column-block-augmented truncations of block-structured Markov chains. Journal of the Operations Research Society of Japan 60, 3 (2017), 271--320.Google Scholar
Cross Ref
- Wouter Minnebo and Benny Van Houdt. 2013. A fair comparison of pull and push strategies in large distributed networks. IEEE/ACM Transactions on Networking 22, 3 (2013), 996--1006.Google Scholar
Digital Library
- Michael Mitzenmacher. 2001. The power of two choices in randomized load balancing. IEEE Transactions on Parallel and Distributed Systems 12, 10 (2001), 1094--1104.Google Scholar
Digital Library
- Brian Munsky and Mustafa Khammash. 2006. The finite state projection algorithm for the solution of the chemical master equation. The Journal of Chemical Physics 124, 4 (2006), 044104.Google Scholar
Cross Ref
- Abhay K Parekh and Robert G Gallager. 1993. A generalized processor sharing approach to flow control in integrated services networks: the single-node case. IEEE/ACM Transactions on Networking 3 (1993), 344--357.Google Scholar
Digital Library
- Abhay K Parekh and Robert G Gallager. 1994. A generalized processor sharing approach to flow control in integrated services networks: The multiple node case. IEEE/ACM Transactions on Networking 2, 2 (1994), 137--150.Google Scholar
Digital Library
- E Seneta. 1967. Finite approximations to infinite non-negative matrices. In Proc. Camb. Phil. Soc, Vol. 63. 983.Google Scholar
Cross Ref
- E Seneta. 1968. Finite approximations to infinite non-negative matrices, II: refinements and applications. In Mathematical Proceedings of the Cambridge Philosophical Society, Vol. 64. Cambridge University Press, 465--470.Google Scholar
Cross Ref
- Abhyudai Singh and Joao Pedro Hespanha. 2006. Lognormal moment closures for biochemical reactions. In IEEE 45th Conference on Decision and Control (CDC). 2063--2068.Google Scholar
Cross Ref
- John N Tsitsiklis and Kuang Xu. 2011. On the power of (even a little) centralization in distributed processing. ACM SIGMETRICS Performance Evaluation Review 39, 1 (2011), 121--132.Google Scholar
Digital Library
- Richard L Tweedie. 1971. Truncation procedures for non-negative matrices. Journal of Applied Probability 8, 2 (1971), 311--320.Google Scholar
Cross Ref
- Richard L Tweedie. 1973. The calculation of limit probabilities for denumerable Markov processes from infinitesimal properties. Journal of Applied Probability (1973), 84--99.Google Scholar
- Benny Van Houdt. 2013. A mean field model for a class of garbage collection algorithms in flash-based solid state drives. ACM SIGMETRICS Performance Evaluation Review 41, 1 (2013), 191--202.Google Scholar
Digital Library
- N. G. Van Kampen. 2007. Stochastic Processes in Physics and Chemistry (3rd ed.). Elsevier.Google Scholar
- Thirupathaiah Vasantam and Ravi R Mazumdar. 2019. Fluctuations Around the Mean-Field for a Large Scale Erlang Loss System Under the SQ (d) Load Balancing. In 2019 31st International Teletraffic Congress (ITC 31). IEEE, 1--9.Google Scholar
Cross Ref
- Qiaomin Xie, Xiaobo Dong, Yi Lu, and Rayadurgam Srikant. 2015. Power of d choices for large-scale bin packing: A loss model. ACM SIGMETRICS Performance Evaluation Review 43, 1 (2015), 321--334.Google Scholar
Digital Library
- Lei Ying. 2016. On the approximation error of mean-field models. ACM SIGMETRICS Performance Evaluation Review 44, 1 (2016), 285--297.Google Scholar
Digital Library
- Lei Ying. 2017. Stein's method for mean field approximations in light and heavy traffic regimes. Proceedings of the ACM on Measurement and Analysis of Computing Systems 1, 1 (2017), 1--27.Google Scholar
Digital Library
- Lulai Zhu, Giuliano Casale, and Iker Perez. 2020. Fluid approximation of closed queueing networks with discriminatory processor sharing. Performance Evaluation 139 (2020), 102094.Google Scholar
Cross Ref
Index Terms
Refining Mean-field Approximations by Dynamic State Truncation
Recommendations
Refining Mean-field Approximations by Dynamic State Truncation
SIGMETRICS '21Mean-field models are an established method to analyze large stochastic systems with N interacting objects by means of simple deterministic equations that are asymptotically correct when N tends to infinity. For finite N, mean-field equations provide an ...
Refining Mean-field Approximations by Dynamic State Truncation
SIGMETRICS '21: Abstract Proceedings of the 2021 ACM SIGMETRICS / International Conference on Measurement and Modeling of Computer SystemsMean-field models are an established method to analyze large stochastic systems with N interacting objects by means of simple deterministic equations that are asymptotically correct when N tends to infinity. For finite N, mean-field equations provide an ...
On the Approximation Error of Mean-Field Models
Performance evaluation reviewMean-field models have been used to study large-scale and complex stochastic systems, such as large-scale data centers and dense wireless networks, using simple deterministic models (dynamical systems). This paper analyzes the approximation error of ...






Comments