China
Africa
Explore chapters and articles related to this topic
Reliability analysis in the presence of Aleatory uncertainty
Published in Stein Haugen, Anne Barros, Coen van Gulijk, Trond Kongsvik, Jan Erik Vinnem, Safety and Reliability – Safe Societies in a Changing World, 2018
L.G. Crespo, S.P. Kenny, D.P. Giesy
Metamodeling (Simpson, Peplinski, Koch, & Allen 2001) refers to the process of creating a mathematical representation of a phenomenon based on input-output data. This paper uses a metamodeling technique for constructing computational models describing the distribution of a continuous output variable. These models are called Random Predictor Models (RPMs) because the predicted output corresponding to any given input is a random variable. One common example of an RPM is a Gaussian Process (GP) model (Rasmussen & Williams 2006). In contrast to GP models, which only lead unimodal and symmetric responses, we focus on RPMs having a bounded support set and prescribed values for the first four moments. The manipulation of these functions enables the generation of predictors that accurately describe possibly skewed and multimodal responses typical of many physical phenomena.
Project Haystack Data Standards
Published in John J. “Jack” Mc Gowan, Energy and Analytics, 2020
The project also provided a more technical explanation of their goals for energy and building data. “The Project Haystack data modeling standard for Buildings and Equipment systems shall use a simple meta-model.” As defined by Wikipedia, “meta-modeling is the analysis, construction and development of the frames, rules, constraints, models and theories applicable and useful for modeling predefined class of problems.” In this case the problem is that data from building systems do not follow consistent rules and are not easily used by applications, hence the need for mapping. As noted, the Project Haystack meta-model is based on the broadly accepted concept of “tags.” To provide more detail on this concept, tags are name/value pairs, associated with entities like AHUs, electric meters, etc. Tags are simple and dynamic, add structure, and provide the flexibility needed to establish standardized models of diverse systems and equipment. Tags are a modeling technique that allows easy customization of data models on a per-task, per-project or per-equipment basis, while retaining the ability to be interpreted by external applications using a standard, defined methodology and vocabulary. Using Haystack, external web-based applications are able to receive data that include essential meta data (tags) to describe the meaning of the data. This approach can be used for graphics, analytics, maintenance management and other applications, and it enables automatic interpretation of the data by software applications. This Project Haystack effort is a gift to the energy and analytics world, and it could only have been done with the expertise of this highly competent community of technologists and entrepreneurs.
DEVS as a Semantic Domain for Programmed Graph Transformation
Published in Gabriel A. Wainer, Pieter J. Mosterman, Discrete-Event Modeling and Simulation, 2018
Eugene Syriani, Hans Vangheluwe
Model-driven approaches are becoming increasingly important in the area of software engineering. In model-driven development, models are constructed to conform to meta-models. A meta-model defines the (possibly infinite) set of all well-formed model instances. As such, a meta-model specifies the syntax and static semantics of models. Meta-models are often described as the Unified Modeling Language (UML) Class Diagrams. In model-driven engineering, meta-modeling goes hand-in-hand with model transformation.
Active learning metamodelling for survival rate analysis of simulated emergency medical systems
Published in Transportmetrica A: Transport Science, 2022
Francisco Antunes, Marco Amorim, Francisco Pereira, Bernardete Ribeiro
Song et al. (2017) summarise the metamodeling methodology in three major steps, namely, (1) definition of the experimental design, (2) metamodel specification, and (3) metamodel learning/fitting. The first step consists of strategically sampling the input space to generate a data set for the metamodel training. Then, steps 2 and 3 are quite intertwined and, for that, they are conducted simultaneously. While the former involves selecting a family of functional forms for the metamodel, the latter regards the fitting procedure of the chosen metamodel to the data set obtained from step 1. For a more in-depth description of the simulation metamodeling approach and unfolding in ten detailed steps suggestions, please refer to (Kleijnen and Sargent 2000).