Insight最新文献

Prevailing modeling and simulation (M&S) techniques have struggled to provide meaningful quantitative results in M&S of complex system of systems (SoSs) in the face of an environment filled with complex interacting uncertainties. This paper reports on systems thinking applied to “how” M&S techniques should shift to allow a next generation of quantitative tools and techniques. The imperative is to provide quantitative performance results across the constituent interfaces in a modeled architecture. A five-step statistical and parametric algorithm tool that addresses uncertainty quantification (UQ) is presented. [Improving the utility of UQ data evaluation] A quantitative approach to managing complex uncertainties across modeled interfaces using graph theory is proposed. A future vision for SoS engineering (SoSE) that uses graph theory-based modeling is suggested to improve the utility of tools such as UQ is suggested.

Uncertainty is a large part of the systems engineering development process. Particularly absent is the quantification of uncertainty of the threat, operating environment, and friendly force factors at each step of this lifecycle. This paper will explore a methodology to quantify the amount of uncertainty and the interdependencies of the uncertainty factors during the development. Included for consideration are internal and external factors and their contribution to the overall system uncertainty. An illustrative example is provided to exercise this methodology.

In early design stages, business developers and systems engineers deal with uncertainties on the business problem, in line with the company's strategy. Before designing the system, the business developers need to set the boundaries of the business problem: What are the values to deliver to which stakeholders? What are their preferences? What are the future trends or the evolution of the markets and the external context? These questions regarding the uncertainties on the definition of the problem may not have answers and need to be investigated to assess the value robustness of the possible design alternatives. The aim of this work is to support decision-making in business and system design thanks to a broad and rapid analysis of a large amount of business design alternatives under uncertainty. We introduce a decision-making support method, called ValXplore, based on visual analysis and data analytics to explore the uncertainties on and in the business problem. The method was tested and validated on an industrial case study to assess the benefits and limits of the semi-reusability of a launch vehicle. Both business developers and systems engineers can rapidly explore a broad space of alternatives to increase the value to the stakeholders, by performing sensitivity and uncertainty analyses.

Although theoretically independent, requirements within a decomposition level of a system architecture are not isolated elements. For an existing design, a change of a requirement may endanger or facilitate fulfillment of other requirements within the same level of the decomposition. The present research suggests a requirement connectivity metric to evaluate the potential consequences that changing a requirement may have on a system with respect to fulfillment of other requirements. A particular aspect of the present research is the assumption that connectivity accounts only for requirements within the same decomposition level of an architecture, not for those flowing up or down the decomposition. The metric is used to evaluate different cases in which requirements are changed due to triggering of uncertain events during a project life cycle.

Although a number of recent studies on using Bayesian Networks (BN) for system reliability estimation have been proposed, these studies are based on the assumption that a pre-built BN was designed to represent the system. In these studies, the task of building the BN is typically left to a group of specialists who are BN and domain experts. However, the process of building a system-specific BN is generally very time consuming and may lead to incorrect deductions. As there are no existing studies to eliminate the need for a human expert in the process of system reliability estimation, this paper introduces a holistic method that uses historical data about the system to be modeled as a BN and provides efficient techniques for automated construction of the BN model and estimation of the system reliability. Moreover, very limited human intervention is sufficient for the process of BN construction and reliability estimation.

Exploratory modelling is an emerging approach which can address the challenge of model-based decision making in dealing with input model uncertainties. Exploratory modelling samples from an input uncertainty space and generates extensive computational experiments to analyse possible model behaviours in an output solution space. The way that the input uncertainty space is delineated influences the results of exploratory modelling and its computational cost. In this article, we show the statistical significance of the implication of the size of an input uncertainty space on the resulted output solution space. We also propose a heuristic approach which informs the delineation of input uncertainties by screening the relevant model behaviour in the solution space. An illustrative example of an aircraft fleet management system is used to demonstrate the implementation of our approach in practice. We conclude that the delineation of input uncertainty space can be a way to control simulations in exploratory modelling and to enhance the efficiency of the exploration process and the confidence of the final results.

Currently, technology readiness assessments (TRAs) are used in determining the maturity of the critical technology elements (CTEs) of a system as it moves forward in the system development life cycle. The TRA method uses technology readiness levels (TRLs) as the decision metric. TRL values are assessed and determined by subject matter experts (SMEs). Since expert evaluators often differ in their judgment when scoring a system element against the TRL scale criteria, this paper argues for the use of a Bayesian network model to provide a mathematical method to consistently combine and validate the judgment of these SMEs and increase the confidence in the determination of the readiness of system components and their technologies.

Traditional systems engineering pays attention to careful composition of prose requirements statements. Even so, prose appears less than what is needed to advance the art of systems engineering into a theoretically based engineering discipline comparable to electrical, mechanical, or chemical engineering. Ask three people to read a set of prose requirements statements, and a universal experience is that there will be three different impressions of their meaning. The rise of model-based systems engineering might suggest the demise of prose requirements, but we argue otherwise. This paper shows how prose requirements can be productively embedded in and a valued formal part of requirements models. This leads to the practice-impacting insight that requirements statements can be non-linear extensions of linear transfer functions, shows how their ambiguity can be further reduced using ordinary language, how their completeness or overlap more easily audited, and how they can be “understood” more completely by engineering tools.