Supervised learning-based approximation method for single-server open queueing networks with correlated interarrival and service times

Tan, Barış; Khayyati, Siamak

doi:10.1080/00207543.2021.1887536

Cited by 8 publications

(5 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The accuracy of the prediction of the output parameters of the Batching, Merge, and Split blocks has been reported in (Tan and Khayyati 2021). These results indicate that GPR yields very accurate predictions for the output parameters for modeling all the blocks considered in this work.…”

Section: Training Datamentioning

confidence: 61%

“…Finally, the output characteristics are fed into the following block to analyze a queueing network. The general method presented in this study is based on the SLQNA algorithm given in (Tan and Khayyati 2021). In Sections 4.1, 4.2, and 4.3, we summarize the steps of SLQNA for completeness.…”

Section: A Supervised Learning Based Prediction Methods For the Outpu...mentioning

confidence: 99%

“…The mean and variance of the prediction can be used to obtain the confidence interval for the prediction. We choose GPR as the main learning method in this work given its ability to model noisy data and its e ciency in modeling similar datasets (Tan and Khayyati 2021).…”

Section: Gaussian Process Regressionmentioning

confidence: 99%

“…The parameters of the initial normal random variables are chosen while considering the changes that will result in the first-lag autocorrelation from the transformations. For the details on the generation of correlated inter-event times and simulating di↵erent building blocks, we refer the reader to (Tan and Khayyati 2021).…”

Section: Generating the Training Data By Simulating The General Delay...mentioning

confidence: 99%

“…The advancement of machine learning tools allows combining the generality of simulation with the computational e ciency of analytical methods. Recently, Tan and Khayyati (2021) presented a supervised learning based approximation method (SLQNA) to analyze open queueing networks with single servers that operate according to First-Come-First-Serve (FCFS) service discipline with correlated interarrival and service times. SLQNA uses local simulations to generate training data for building supervised learning blocks that approximate the behavior of the various blocks in the queueing network.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Supervised-learning-based approximation method for multi-server queueing networks under different service disciplines with correlated interarrival and service times

Khayyati

Tan

2021

International Journal of Production Research

Self Cite

View full text Add to dashboard Cite

Developing e cient performance evaluation methods is important to design and control complex production systems e↵ectively. We present an approximation method (SLQNA) to predict the performance measures of queueing networks composed of multi-server stations operating under di↵erent service disciplines with correlated interarrival and service times with merge, split, and batching blocks separated with infinite capacity bu↵ers. SLQNA yields the mean, coe cient of variation, and first-lag autocorrelation of the inter-departure times and the distribution of the time spent in the block, referred as the cycle time at each block. The method generates the training data by simulating di↵erent blocks for di↵erent parameters and uses Gaussian Process Regression to predict the inter-departure time and the cycle time distribution characteristics of each block in isolation. The predictions obtained for one block are fed into the next block in the network. The cycle time distributions of the blocks are used to approximate the distribution of the total time spent in the network (total cycle time). This approach eliminates the need to generate new data and train new models for each given network. We present SLQNA as a versatile, accurate, and e cient method to evaluate the cycle time distribution and other performance measures in queueing networks.

show abstract

Section: Training Datamentioning

confidence: 61%

Section: A Supervised Learning Based Prediction Methods For the Outpu...mentioning

confidence: 99%

Section: Gaussian Process Regressionmentioning

confidence: 99%

Section: Generating the Training Data By Simulating The General Delay...mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Supervised-learning-based approximation method for multi-server queueing networks under different service disciplines with correlated interarrival and service times

Khayyati

Tan

2021

International Journal of Production Research

Self Cite

View full text Add to dashboard Cite

show abstract

On general multi-server queues with non-poisson arrivals and medium traffic: a new approximation and a COVID-19 ventilator case study

Gosavi

2022

Oper Res Int J

View full text Add to dashboard Cite

We consider the multi-server, single-channel queue, i.e., a G / G / k queue with k identical servers in parallel, under the first-come-first-served discipline in which the inter-arrival process is non-Poisson, the service time has any given distribution, and traffic is of medium intensity. Such queues are common in factories, airports, and hospitals, where the inter-arrival times and service times are typically not exponentially distributed, but rather have double-tapering distributions whose probability density functions taper on both sides, e.g., gamma, triangular etc. For these conditions, a new closed-form approximation based on only the mean and variance of the two inputs, the inter-arrival and service times, is presented. Determining distributions of inputs typically requires additional human effort in terms of histogram-fitting and running a goodness-of-fit test, which is avoided here. The new approximation is tested on a variety of scenarios and its performance is benchmarked against simulation. Further, the new approximation is also implemented on a ventilator case study from the recent COVID-19 pandemic to demonstrate its utility in optimizing server capacity. The approximation provides errors typically in the range 1–15% and 31% in the worst case. In systems where data change rapidly and decisions must be made quickly, this approximation will be particularly useful.

show abstract

Using Recurrent Neural Networks for the Performance Analysis and Optimization of Stochastic Milkrun-Supplied Flow Lines

Südbeck,

Mindlina,

Schnabel

et al. 2024

Schmalenbach J Bus Res

View full text Add to dashboard Cite

Long-term throughput, as a key performance indicator of a stochastic flow line, is affected by numerous parameters describing the features of the flow line, such as processing time and buffer size. Fast and accurate evaluation methods for a given set of values for those parameters are a prerequisite to systematically optimize such a flow line. In this paper, we consider the case of a flow line with random processing times, limited buffer capacities and so-called milkruns that supply the machines with material parts that are required to perform, e.g., assembly operations on workpieces. In such a system, shortages in the supply of material parts can limit the performance of the flow line. Up to now, there are no accurate analytical approaches to quantify the complex interactions in such milkrun-supplied flow lines for realistic problem sizes. We propose to use recurrent neural networks to determine the long-term throughput of such flow lines enabling us to evaluate production systems of flexible size. Our results show that the throughput can be determined accurately and quickly via recurrent neural networks. Furthermore, we use this new evaluation procedure as a building block to optimize this type of flow line using gradient and local search techniques.

show abstract

Supervised learning-based approximation method for single-server open queueing networks with correlated interarrival and service times

Cited by 8 publications

References 30 publications

Supervised-learning-based approximation method for multi-server queueing networks under different service disciplines with correlated interarrival and service times

Supervised-learning-based approximation method for multi-server queueing networks under different service disciplines with correlated interarrival and service times

On general multi-server queues with non-poisson arrivals and medium traffic: a new approximation and a COVID-19 ventilator case study

Using Recurrent Neural Networks for the Performance Analysis and Optimization of Stochastic Milkrun-Supplied Flow Lines

Contact Info

Product

Resources

About