A Bayesian Approach to Sequential Optimization based on Computer Experiments

Tajbakhsh, Sam Davanloo; Rosenberger, James L.

doi:10.1002/qre.1658

Cited by 10 publications

(7 citation statements)

References 30 publications

(40 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Meanwhile, the distinction is made between the risks associated with model-based decision making (Olson, 2015), where models depend on the current market situation and the risks of production activity. The use of probabilistic models is based on the use of risk assessments (Mylnikov and Kuetz, 2017), the Bayes theorem (Tajbakhsh et al, 2015) and the Monte Carlo method (Moghaddam, 2015). In the field of risk assessment, the significant contribution of Markowitz (1952), Mossin (1961), Sharpe (1964) and Lintner (1966) should also be mentioned.…”

Section: Literature Reviewmentioning

confidence: 99%

Production Planning with Parameters on the Basis of Dynamic Predictive Models: Interconnection and the Inertness of their Interaction

Mylnikov¹,

Fayzrakhmanov²

2018

ERSJ

View full text Add to dashboard Cite

The research is related to the increasing role of prognostic models in production systems management, which is associated with an increase in the requirements for managerial efficiency, the need to consider external factors affecting the system, the determination of the features of the systems in question, the examination of the processes in progress and the relationship between the chain of managerial decisions and the values of the selected control parameters. The purpose of the article is to consider and evaluate the consequences of decisions made as a chain of interrelated events in time with regard to the dynamics of the environment in which production systems operate and the variability of control parameters. The leading approach of the research considers the production system as one that is open "in terms of environment" and "in terms of the ultimate goal". The proprietary results demonstrate that the solutions obtained are of a probabilistic nature, the solutions should be set by ranges of possible values, the decision ranges can be arranged in such a way as to introduce variability into the decisions made, the choice of which will be based on factors not taken into account in the proposed method of analyzing production systems. The practical and theoretical significance of the research is that the described methodology allows to obtain optimal values of control parameters based on the objectives of the production system under consideration on the basis of its integrated assessment, taking into account the interaction and the mutual influence of the system's parameters, their inertness and probabilistic nature, which makes it possible to increase the validity of managerial decisions and to consider the inertness of the processes taking place in the system during planning.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Production Planning with Parameters on the Basis of Dynamic Predictive Models: Interconnection and the Inertness of their Interaction

Mylnikov¹,

Fayzrakhmanov²

2018

ERSJ

View full text Add to dashboard Cite

show abstract

“…Probabilistic models involve using risk assessment (Abdullaev et al, 2012), the Bayes theorem (Tajbakhsh et al, 2015) or the Monte Carlo method (Moghaddam, 2015). Such approaches allow moving from risk assessment of individual cases and tasks to the consideration of projects, processes and the PS as a whole.…”

Section: Literature Overviewmentioning

confidence: 99%

The Risk Assessment Method in Prognostic Models of Production Systems Management with Account of the Time Factor

Mylnikov¹,

Kuetz²

2017

ERSJ

View full text Add to dashboard Cite

The possibility of planning risks assessment and the setting of the planning horizon are analysed with account of the time factor basing. A risk metrics approach is used, with risks estimated as a time function given in tabular form for the purposes of master production. Dependence of the integrated planning risks on the accuracy of forecasts used is proved and a time point is identified after which the risk assessment value increases sharply. The study enables to obtain the values of planning risks integral assessment in management problems using forecast data for groups of indicators and parameters involved in decisionmaking.

show abstract

“…This dilemma makes it difficult for engineers to run the computer code too many times. Hence, from the statistics perspective, engineers need to construct a "cheap" model just based on several runs of the complex computer model/code 2 . Engineers can replace the time‐consuming complex computer model/code with the "cheap" model to solve the optimization problem.…”

Section: Introductionmentioning

confidence: 99%

“…Hence, from the statistics perspective, engineers need to construct a "cheap" model just based on several runs of the complex computer model/code. 2 Engineers can replace the time-consuming complex computer model/code with the "cheap" model to solve the optimization problem. This "cheap" model is usually regarded as a metamodel or surrogate.…”

Section: Introductionmentioning

confidence: 99%

Incremental Kriging based online robust parameter design

Zhou

Wang

Jiang

et al. 2022

Quality & Reliability Eng

View full text Add to dashboard Cite

Traditional robust parameter design (RPD) is a useful tool for offline quality improvement, which is mainly applied to physical experiment. If the optimal setting for controllable factors is not desirable, the offline PRD cannot utilize the new‐obtained sample to update the optimal setting. In this paper, the offline RPD is extended to online PRD; and the background for RPD is enlarged from physical experiment to computer experiment, where the Kriging model is typically utilized to create a single response surface. In the process of building the incremental Kriging (IKriging), if the optimal setting is not desirable, it can be regarded as a new‐arrival sample with the online measurements of the noise factors to update the Kriging model. While constructing the response surface with the IKriging, it is updated according to the new‐arrival sample rather than the all (n+1) samples, substantially enhancing the efficiency of constructing the response surface. In the proposed IKriging based online RPD strategy, an appropriate algorithm is designed to numerically get the next optimal setting for the controllable factors, and a criterion is established to determine the next site for the noise factors to observe the response. By employing the IKriging model, the response surface can be online constructed efficiently, and furthermore the optimal setting for the controllable factor can be obtained immediately, meeting the requirement of online quality adjustment. The proposed methodology is confirmed using five examples, which illustrate that the underlying optimal settings can be found by utilizing the proposed IKriging based online RPD strategy for the computer experiment, and the IKriging based RPD method is preferable to existing online robust design methods.

show abstract

A Bayesian Approach to Sequential Optimization based on Computer Experiments

Cited by 10 publications

References 30 publications

Production Planning with Parameters on the Basis of Dynamic Predictive Models: Interconnection and the Inertness of their Interaction

Production Planning with Parameters on the Basis of Dynamic Predictive Models: Interconnection and the Inertness of their Interaction

The Risk Assessment Method in Prognostic Models of Production Systems Management with Account of the Time Factor

Incremental Kriging based online robust parameter design

Contact Info

Product

Resources

About