History of input modeling

Cheng, Russell C. H.

doi:10.1109/wsc.2017.8247789

Cited by 12 publications

(12 citation statements)

References 79 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…General Electric has been developing generic auto-simulation capability to provide data-driven simulation capability targeting manufacturing applications (Annunziata and Biller 2012). Cheng and Law (2017) discuss data-driven simulations in the present age of big data, providing a summary of work that has been presented primarily at the Winter Simulation Conference over the years.…”

Section: Operating Policy Learning Via Simulation Optimizationmentioning

confidence: 99%

The role of learning on industrial simulation design and analysis

Biller¹,

Biller²,

Dülgeroğlu³

et al. 2017

2017 Winter Simulation Conference (WSC)

View full text Add to dashboard Cite

The capability of modeling real-world system operations has turned simulation into an indispensable problemsolving methodology for business system design and analysis. Today, simulation supports decisions ranging from sourcing to operations to finance, starting at the strategic level and proceeding towards tactical and operational levels of decision-making. In such a dynamic setting, the practice of simulation goes beyond being a static problem-solving exercise and requires integration with learning. This article discusses the role of learning in simulation design and analysis motivated by the needs of industrial problems and describes how selected tools of statistical learning can be utilized for this purpose.

show abstract

Section: Operating Policy Learning Via Simulation Optimizationmentioning

confidence: 99%

The role of learning on industrial simulation design and analysis

Biller¹,

Biller²,

Dülgeroğlu³

et al. 2017

2017 Winter Simulation Conference (WSC)

View full text Add to dashboard Cite

show abstract

“…Literature reviews show that data gathering and input modeling is an active field in simulation modeling (see [8]). While more technical introductionary texts on simulation mention data gathering only in passing (see [3] and [30]), Carson (see [5]) describes in his overview on the modeling and simulation process often encountered challenges with data collection, cleansing and analysis: data might not be available or not of the desired quality, unforeseen effort is necessary to clean up databases and files, with clients often not knowing what data they own and whether it is acurate.…”

Section: Introductionmentioning

confidence: 99%

Co-operating with Municipal Partners on Indicator Identification and Data Acquisition

Ullrich¹,

Bogen²,

Lückerath³

et al. 2019

SNE

View full text Add to dashboard Cite

Data acquisition is a significant part of the simulation modeling process. It can be a challenge, especially in larger projects with many participants and stakeholders from different professions: They all bring their individual understanding of data, its availability, formats, and quality, with their respective points of view strongly dependent on their own professional backgrounds. This paper presents a modular process for indicator identification and data acquisition in the context of modeling projects, specifically aimed at researchers or practitioners working with municipal partners, i.e. representatives from city councils, urban planners and administrations, or infrastructure providers. The process consists of five steps: indicator selection, suitability check, data gathering, data quality check, and data management. To further facilitate a clearer understanding of the endeavor, the process defines necessary input and to be expected output artifacts. In addition to presenting the process, the paper describes its examplary execution with partners from the City of Bilbao, Spain, and offers a selection of discovered challenges and lessons learned.

show abstract

“…Input modelling for discrete-event simulation (DES) aims to identify appropriate probability distributions for characterising the behaviour of the streams of random variables that represent the inputs to DES models. A recent review of input modelling has been given by Cheng (2017a) showing how the topic has grown so that there is now an extensive literature. The most basic situation is the simple one where input random variables are independently and identically distributed and drawn from wellknown distributions such as the normal, lognormal, gamma or Weibull.…”

Section: Introductionmentioning

confidence: 99%

“…A wider range of distributional shapes has been discussed in Kuhl et al (2010). Two generalizations, reviewed by Cheng (2017a), have been studied in some detail, namely: (i) where the random variables are multivariate, and (ii) where they are correlated. See for example Deler and Nelson (2001); Ghosh and Henderson (2001); Nelson and Yamnitsky (1998).…”

Section: Introductionmentioning

confidence: 99%

“…See for example Deler and Nelson (2001); Ghosh and Henderson (2001); Nelson and Yamnitsky (1998). A third generalization, though mentioned in Cheng (2017a) and Kuhl et al (2010), has not been so well discussed, where input random variables have a multimodal distribution, and most likely follow a finite mixture distribution. The purpose of this article is to discuss such distributions and their modelling in DES.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Input modelling for multimodal data

Cheng

Currie

2019

Journal of the Operational Research Society

View full text Add to dashboard Cite

Multimodal data occurs frequently in discrete-event simulation input analysis, typically arising when an input sample stream comes from different sources. A finite mixture distribution is a simple input model for representing such data, but fitting a mixture distribution is not straightforward as the problem is well-known to be statistically non-standard. Even though much studied, the most common fitting approach, Bayesian reversible jump Markov Chain Monte Carlo (RJMCMC), is not very satisfactory for use in setting up input models. We describe an alternative Bayesian approach, MAPIS, which uses maximum a posteriori estimation with importance sampling, showing it overcomes the main problems encountered with RJMCMC. We demonstrate use of a publicly-available implementation of MAPIS, which we have called FineMix, applying it to practical examples coming from finance and manufacturing.

show abstract

History of input modeling

Cited by 12 publications

References 79 publications

The role of learning on industrial simulation design and analysis

The role of learning on industrial simulation design and analysis

Co-operating with Municipal Partners on Indicator Identification and Data Acquisition

Input modelling for multimodal data

Contact Info

Product

Resources

About