Development Of Metamodels Research Articles

N -Dimensional Nonuniform Rational B-Splines for Metamodeling

Nonuniform rational B-splines (NURBs) have unique properties that make them attractive for engineering metamodeling applications. NURBs are known to accurately model many different continuous curve and surface topologies in one- and two-variate spaces. However, engineering metamodels of the design space often require hypervariate representations of multidimensional outputs. In essence, design space metamodels are hyperdimensional constructs with a dimensionality determined by their input and output variables. To use NURBs as the basis for a metamodel in a hyperdimensional space, traditional geometric fitting techniques must be adapted to hypervariate and hyperdimensional spaces composed of both continuous and discontinuous variable types. In this paper, we describe the necessary adaptations for the development of a NURBs-based metamodel called a hyperdimensional performance model or HyPerModel. HyPerModels are capable of accurately and reliably modeling nonlinear hyperdimensional objects defined by both continuous and discontinuous variables of a wide variety of topologies, such as those that define typical engineering design spaces. We demonstrate this ability by successfully generating accurate HyPerModels of ten trial functions laying the foundation for future work with N-dimensional NURBs in design space applications.

Nonlinear regression metamodels: a systematic approach

This paper proposes an approach for systematic development of nonlinear regression metamodels for stochastic simulation. This approach provides the practitioner with a process for the construction of nonlinear metamodels in general, and includes statistical techniques for estimation and validation of nonlinear regression models. In order to ensure that the resulting metamodel is a valid substitute for the original simulation model, validation techniques are suggested. In a case study, the proposed application leads to simple function that adequately approximate the model's behaviour, while linear regression polynomials result in a poor fit.

Simulation metamodel development using uniform design and neural networks for automated material handling systems in semiconductor wafer fabrication

Simulation is very time consuming, especially for complex and large scale manufacturing systems. The process of collecting adequate sample data places limitations on any analysis. This paper proposes to overcome the problem by developing a neural network simulation metamodel that requires only a comparably small training data set. In the training data set, the configuration of all input data is generated by uniform design and the corresponding output data are the result of simulation runs. A dispatching problem for a complex simulation model of an automated material handling system (AMHS) in semiconductor manufacturing is introduced as an example. In the example, there are 23 4-levels factors, resulting in a total of 4 23 possible configurations. However, by using the method proposed in this paper, only 28 configurations had to be simulated in order to collect the training data. The results show that the average prediction error was 3.12%. The proposed simulation metamodel is efficient and effective in solving a practical application.

Multidimensional sequential sampling for NURBs-based metamodel development

Adaptive design of experiments approaches are intended to overcome the limitations of a priori experimental design by adapting to the results of prior runs so that subsequent runs yield more significant information. Such approaches are valuable in engineering applications with metamodels, where efficiently collecting a dataset to define an unknown function is important. While a variety of approaches have been proposed, most techniques are limited to sampling for only one phenomenon at a time. We propose a multicriteria optimization approach that effectively simultaneously samples for multiple phenomena. In addition to determining the next sequential sampling point, such an algorithm also can be formulated to support conclusions about the adequacy of the experiment through the use of convergence criteria. A multicriteria adaptive sequential sampling algorithm, along with convergence metrics, is defined and demonstrated on five trial problems of engineering interest. The results of these five problems demonstrate that a multicriteria sequential sampling approach is a useful engineering tool for modeling engineering design spaces using NURBs-based metamodels.

An Artificial Neural Network Based Metamodel for Analysing a Stochastic Combat Simulation

In this paper, we discuss the use of multi-layer perceptron (MLP) neural networks to predict the outputs from a stochastic combat simulation model. The use of a faster, auxiliary model to approximate the input-output relationships of a complex simulation model is known in the simulation community as metamodelling. Since many simulation models contain a large number of input parameters, however it is necessary to reduce the set considered and determine the most important ones to include in a metamodel given a particular modelling context. This paper employs a two-stage experimental design — first, Morris’ randomised one-at-a-time (OAT) design is used as a factor screening method, and then a Latin Hypercube design is employed to select the set of input configurations which defines the required runs of the simulation model. The set of input configurations together with the associated simulation outputs provide the training data for the development of the MLP networks. The approach is illustrated with reference to a stochastic combat simulation called SIMBAT. The paper then investigates a number of aspects relating to the development of neural network-based metamodels of stochastic combat simulations. It shows that using the outputs from each replication of a stochastic simulation is generally better than only using the mean output.

아키텍처산출물 관리 시스템 개발

현재 국방부에서는 국방정보체계간 상호운용성을 보장하기 위해 아키텍처산출물을 쉽고 일관성이 있게 개발할 수 있는 국방아키텍처프레임워크와 개발된 아키텍처산출물을 저장하는 핵심아키텍처데이터모델을 개발하였다. 하지만 작성된 아키텍처산출물을 문서로 관리하기에는 많은 어려움이 존재함으로, 아키텍처를 모델링할 수 있고 모델링된 정보를 공통의 저장소에 관리할 수 있는 모델링 도구 및 리포지토리 시스템이 필요하다. 본 논문에서는 국방아키텍처프레임워크와 핵심아키텍처데이터모델의 메타모델과 아키텍처산출물 작성 및 관리를 지원하는 아키텍처산출물 관리 시스템을 개발하였고 기존 시스템과 비교 평가하였다. 이를 통해 각 기관의 아키텍처개발자가 아키텍처산출물을 모델링 방법으로 보다 쉽고 일관성 있게 작성하고 관리할 수 있으며, 다양한 질의를 통해 시스템에서 관리되는 아키텍처산출물 정보를 유용하게 검색 및 조회하여 의사결정에 참조할 수 있다. 또한 작성된 아키텍처산출물을 통합하고 분석 및 비교하여 정보체계간 상호운용성과 비용 절감을 지원한다. MND(Ministry of National Defense) has developed MND AF(Ministry of National Defense Architecture Framework) and CADM(Core Architecture Data Model to guarantee interoperability among defense information systems. But, it is very difficult to manage architecture product documented through MND AF and CADM. So, there Is necessity for development of modeling tool and repository system which can develop architecture products and manage architecture product informations in common repository In this paper, we developed architecture product management system which supports development and management of meta model and architecture product of MND AF and CADM. Through architecture product management system architect of each agency can construct architecture product in a more effective and efficient way with modeling method and a user can search and refer useful architecture product informations using query function. Also, architecture product management system provides the basis for system integration and interoperability with integration, analysis and comparison of architecture product.

FACILITATING MEASUREMENT INDICATORS IN SOFTWARE IMPROVEMENTS PROJECTS

This paper reports on the design and development of a Measurement Metamodel to facilitate the Implementation and Monitoring of Software Improvement initiatives. The Indicators provided for the Measurement Metamodel are building on a Data Warehouse environment. The Measurement Metamodel is designed based on the International Standard ISO 15939 – Software Measurement Process [1]. The Measurement Metamodel supports Practical Software Measurement (PSM) [2] approach and Capability Maturity Model Integrated (CMMI) [3] implementations. The Measurement Metamodel is based on the principle that Software measurement process must be flexible and tailorable to the particular information needs of decision makers.

Response surface analysis of a multi-product batch processing facility using a simulation metamodel

In a multi-product batch production environment, managers and engineers must have a robust understanding of the impact of both operational and environmental variables and their interactions on process throughput. This insight is never more important than when a new facility is designed and built. In this study of a particular company in the specialty chemicals industry, a general simulation metamodel was developed in order to understand better how plant design parameters could be optimized in order to maximize plant throughput under certain environmental conditions and with certain asset investment constraints. The simulation metamodel developed is essentially a deterministic, analytical representation of a detailed, discrete-event, stochastic simulation model that was previously developed and validated to study operational and scheduling issues. The development of the simulation metamodel for plant design, as a function of capacity and throughput, is described in this paper. This study exemplifies the application of detailed, discrete-event, stochastic simulation models to derive deterministic, analytical response surface representations for purposes of plant design and throughput analysis in the process industries.

Towards a general web rule language

A general web rule (markup) language has several purposes. It may serve as a lingua franca to exchange rules between different systems and tools. It may be used to express derivation rules for enriching web ontologies by adding definitions of derived concepts or for defining data access permissions; to describe and publish the reactive behaviour of a system in the form of reaction rules; and to provide a complete XML-based specification of a software agent. Further uses may arise in novel web applications. In this paper, we consider the problem of how to design a general web rule language that can be used for these and for future emerging purposes. Given the great diversity of rule concepts and existing rule languages, such a language will consist of several overlapping sublanguages that share a common metamodel. The development of this rule metamodel is a difficult conceptualisation and integration problem.

A comparison of experimental designs in the development of a neural network simulation metamodel

In this case study, we investigate the effects of experimental design on the development of artificial neural networks as simulation metamodels. A simple, deterministic combat model developed within the paradigm of system dynamics provides the underlying simulation. The neural network metamodels are developed using six different experimental design approaches. These include a traditional full factorial design, a random sampling design, a central composite design, a modified Latin Hypercube design and designs supplemented with domain knowledge. The results from this case study show how much impact the experimental design chosen for the neural network training set can have on the predictive accuracy achieved by the metamodel. We compare the networks in terms of various performance measures. The neural network developed from the modified Latin Hypercube design supplemented with domain knowledge produces the best performance, outperforming networks developed from other designs of the same size.

A Meta-Model for Fact Extraction from Delphi Source Code

Fact extraction by parsing is often the first step when analyzing a software system in a reverse engineering context. Essential to a fact extractor is the underlying meta-model, which specifies the elements and relations to be extracted. In this work we will introduce a meta-model for the Delphi programming language. The meta-model will be compared against the Dagstuhl middle model (DMM) and reasons for the development of an additional meta-model will be given. Furthermore, we will report on our first experiences with a fact extractor currently under development. We evaluate our fact extractor with two software applications that were developed in the open source community. In particular we give numbers and examples to point out the capabilities and the not yet resolved open issues of our fact extractor and we will reflect our experiences made during the case studies.

Simulation metamodeling through artificial neural networks

Simulation metamodeling has been a major research field during the last decade. The main objective has been to provide robust, fast decision support aids to enhance the overall effectiveness of decision-making processes. This paper discusses the importance of simulation metamodeling through artificial neural networks (ANNs), and provides general guidelines for the development of ANN-based simulation metamodels. Such guidelines were successfully applied in the development of two ANNs trained to estimate the manufacturing lead times (MLT) for orders simultaneously processed in a four-machine job shop. The design of intelligent systems such as ANNs may help to avoid some of the drawbacks of traditional computer simulation. Metamodels offer significant advantages regarding time consumption and simplicity to evaluate multi-criteria situations. Their operation is notoriously fast compared to the time required to operate conventional simulation packages.

Simplifying Sirius: sensitivity analysis and development of a meta-model for wheat yield prediction

A sensitivity analysis and analysis of the structure of the Sirius wheat model has resulted in the development of a simpler meta-model, which produced very similar yield predictions to Sirius of potential and water-limited yields at two locations in the UK, Rothamsted and Edinburgh. This greatly increases the understanding of the nature and consequences of the relationships implicit within Sirius. The analysis showed that the response of wheat crops to climate could be explained using a few simple relationships. The meta-model aggregates the three main Sirius components, the calculation of leaf area index, the soil water balance model and the evapotranspiration calculations, into simpler equations. This results in a requirement for calibration of fewer model parameters and means that weather variables can be provided on a monthly rather than a daily time-step, because the meta-model can use cumulative values of weather variables. Consequently the meta-model is a valuable tool for regional impact assessments when detailed input data are usually not available. Because the meta-model was developed from the analysis of Sirius, rather than from statistical fitting of yield to weather data, it should perform well for other locations in Great Britain and with different management scenarios.

Dealing with the least squares regression assumptions in simulation metamodeling

The aim of this study is twofold. The first is to estimate a metamodel for a time-shared computer system using a sequential design procedure. The second is to deal extensively with the least squares regression assumptions during the metamodel development. In the first stage of the experimentation, a first-order metamodel is estimated using the two-level factorial design. Later, the design is augmented with replicated center points for curvature check. Upon the detection of the significance of the curvature, a central composite design is used for fitting a second-order metamodel, which explains the relation between the levels of the input factors, and the response of interest. In both stages, various diagnostic statistical tests such as normality test, variance homogeneity test, lack-of-fit test and etc. are carried out to make sure that the method of least squares is properly and efficiently applied.

A comparison of factorial and random experimental design methods for the development of regression and neural network simulation metamodels

This paper compares two forms of experimental design methods that may be used for the development of regression and neural network simulation metamodels. The experimental designs considered are full factorial designs and random designs. The paper shows that, for two example problems, neural network metamodels using a randomised experimental design produce more accurate and efficient metamodels than those produced by similar sized factorial designs with either regression or neural networks. The metamodelling techniques are compared by their ability to predict the results from two manufacturing systems that have different levels of complexity. The results of the comparison suggest that neural network metamodels outperform conventional regression metamodels, especially when data sets based on randomised simulation experimental designs are used to produce the metamodels rather than data sets from similar sized full factorial experimental designs.

A Comparison of Factorial and Random Experimental Design Methdos for the Development of Regression and Neural Network Simulation Metamodels

A Comparison of Factorial and Random Experimental Design Methdos for the Development of Regression and Neural Network Simulation Metamodels

Multivariate regression metamodel: A DSS application in industry

A materials handling system simulation (written using GPSS/H) was developed to predict the Automated Guided Vehicle requirements necessary for a major manufacturer to maintain desired levels of production in one of its automobile assembly plants. Rather than use the simulation as a representational DSS and risk complicating the user interface, validated simulation outputs were collected and used to produce a multivariate regression metamodel. This metamodel formed the centerpiece of a narrow-scope suggestion model DSS used on the factory floor to aid in day to day allocations of resources. This article looks at the metamodel development methodology and offers this technique as an effective means of producing a suggestion model DSS from a more complex representational DSS.

A Metamodeling Approach to Evaluate Agricultural Policy Impact on Soil Degradation in Western Canada

A novel approach for integrating economic and environmental models is described in the context of evaluating soil degradation impacts of agricultural policy in western Canada. The key element of this approach is the development of metamodels, which are statistical summary functions of simulation data obtained from carefully designed experiments with physical process models. The metamodels are in turn used to predict the soil degradation impacts of farmers'land management responses to policy options. The metamodels provide flexibility to perform repeated policy scenarios without having to rerun the time‐and resource‐consuming physical process simulation models. The estimated wind and water erosion metamodels are very robust, with the majority possessing R‐square values in the range of 0·80 to 0·97. The efficiency of the metamodels in facilitating the integration of a policy modeling system is described and applied to a scenario of increased crop residue management. Using regional aggregates of net farm income, total economic, surplus (consumer plus producer surplus) and total soil loss the economic and environmental tradeoff between the status quo and a no‐till policy scenario is evaluated. The model‐predicted economic welfare and environmental quality interaction suggests a clear win‐win situation for society under this alternative policy scenario.

Estimating simulation metamodel parameters for unexpected shop floor real time events

Abstract In order to better understand assembly line system dynamic behavior and provide a basis for developing effective production controls, three unexpected real time events are studied, including sudden machine breakdown at a work station, parts supply shortage and high priority job order processing. Short term system performance is evaluated by dynamic response variables associated with assembly time and work in process (WIP) inventory. Simulation is used as the modeling tool, and the results are further described by mathematical metamodels for overall system behavior. The system transients can be adequately represented by first order exponential functions, and satisfactory results are also obtained for expressing simultaneous occurrence of dynamic events by combined metamodels using linear additivity. A procedure is proposed for the development of dynamic metamodels. For a given system, only a limited number of simulation runs should be made in advance to estimate the metamodel parameters. Using the resultant application program that only contains the metamodels, whenever an event occurs, without running simulation a user can instantaneously predict the consequence in a specified time period by an estimation to the number of assembly products to be finished late.

Evaluating multiprocessor performance: an exercise in metamodeling

This paper points to the viability of metamodels in performance evaluation. A metamodel, in the sense it is used here, is a simple mathematical expression that describes a performance measure in terms of the system parameters, and is derived from experi ments with a simulation model of the system. The system con sidered is a bus-sufficient multiprocessor, and the performance measure under focus is its processing efficiency. The process of metamodel development is illustrated to show that a reliable metamodel can be derived with relative ease. It is also argued that with sufficient care taken in selecting the form of the meta model, it may be justifiable to use the metamodel beyond the experimental domain over which it is developed. Specifically, a metamodel is proposed for the processing efficiency of a multi processor system. It is shown to be a reasonably good approxima tion to the simulation model, and by extension, to the original system.