P-SEMP: A platform for systems engineering modeling and planning
Reilley, Kevin A.
MetadataShow full item record
Systems engineering management and planning has long been a realm dominated by arcane standards, by the weight of years of practice, and by authority. However, with technological advances and the desire to solve socio-technical problems at the level of increasingly complex systems, authority alone is no longer sufficient for the justification of systems engineering practice. As new systems engineering methodologies are bought and sold in the transition towards model-based systems engineering, there is an imperative for the systems engineering practitioner to develop new techniques for estimating project performance before project completion. That is, whether debating appropriate corrective actions for a project at risk of going over budget or over schedule, or when planning a new systems engineering methodology, the systems engineer must forecast planned performance of systems engineering tasks. To this end, the International Council of Systems Engineers (INCOSE) and others have sought to bolster systems engineering measurement and the development of standardized leading indicators of systems engineering performance, which are thought to give insight into future performance in the course of program performance. Recent efforts have produced models of systems engineering performance; however, no model is yet sufficient for addressing which tasks in support of standardized processes should be planned in a systems engineering methodology. This document lays out how such a capability might be implemented by a platform for the numerical comparison of systems engineering methodologies. The idea of a platform for systems engineering modeling and planning is called P-SEMP. There are two threads in this document: a thesis and a methodology. First and foremost, the document is a thesis. The thesis, called at times the P-SEMP Thesis, is a formal argument as to how to address the problem of systems engineering task planning constructed on the basis of gaps, research questions, hypotheses, experiments, and their results. The P-SEMP Thesis aims to prove the best means for determining which systems engineering methodologies, and in particular which methods for a given systems engineering process, are better or worse. Enabling the argument of the P-SEMP Thesis is the P-SEMP Methodology, which is rooted in the fundamentals of modeling and simulation theory but made specific to the class of problems involved in systems engineering methodology comparison. The P-SEMP Methodology describes how to build a platform for P-SEMP and what a platform may entail, and the methodology is supported by a conceptual architecture description. The combined product of the P-SEMP Methodology and conceptual architecture description is a recipe: first, a recipe in terms of the proposed experiments, and then a recipe for the experimental results and conclusions of the P-SEMP Thesis and how its findings may be further applied. In order to render the P-SEMP Thesis manageable in scope, the focus will be placed on tasks surrounding the systems engineering process of validation. Validation, in different senses, can occur both early and late in the system life cycle. While validation is a controversial term, many authors agree that efforts around feasibility assessment, requirements quantification, and the early evaluation of system architectures and design against these requirements are crucial steps in early-phase validation to ensure that the system will meet stakeholder expectations before proceeding with the entirety of the system lifecycle. Concretely, as proposed sets of tasks, a portion of an Object-Oriented Systems Engineering Methodology-inspired process for Spacecraft Requirements Derivation is compared against the State Analysis Model Development method, and subsequently a third method is proposed as well regarding validation concepts. These methods for validation will be modeled and compared using the tooling developed in support of the argument for a platform for systems engineering modeling and planning, the P-SEMP Thesis, and be constructed according to the P-SEMP Methodology with results as shown in the conceptual architecture description for Platform 0.1 and Platform 1.0. The result of the experimental efforts culminates, in a concrete sense, with a domain-specific language for describing tasks in a manner suitable for simulation of the method models. However, leading indicator models are not forgotten; one in particular is replicated and added to a system modeling environment alongside the method models --- however, serious issues in parameterization are uncovered in these leading indicator models and they may not provide much insight towards task planning. Due to these issues and more, a hybrid model proved infeasible in the current situation, leading to the evolution from conceptual Platform 0.1 to the final Platform 1.0. Additionally, as the Spacecraft Requirements Derivation method is proposed specifically for a canonical system FireSat, specific modeling practice in SysML will be proposed to represent the third proposed SE methodology being compared, which requires representation of designs of experiments and probability distributions in the course of ensuring system feasibility. Another motivation for incorporating these expressions into a system model is to ensure the correctness of analytical models which underlay validation processes. This correctness is established by model verification and validation. As these analytical models represent the system from different perspectives, it is beneficial for them to be closely coupled to a unified system model depiction. However, a gap exists where while such capability is known for Multidisciplinary Design Optimization (MDO) and system models, it does not yet exist for Robust Design Simulation (RDS) or techniques for probabilistic or uncertain design processes in conjunction with a system model. Such a technology helps to support the activities above and improves confidence in the results of the early-phase system validation actions. In summation, according to the argument of the P-SEMP Thesis and the practice of the P-SEMP Methodology, a leading indicator model is replicated and found wanting. Systems engineering method model simulations are formulated, and a domain-specific language is created to capture them in the system model for exploration of task architecture. Finally, broader description of designs of experiments and probability are incorporated to improve analytical integration capabilities required for full validation activities in support of greater systems engineering methodology capability. Synthesizing the experimental results is the P-SEMP conceptual architecture Platform 1.0, which serves as a new baseline for systems engineering task planning and comparison, and which places the results into the greater context of how to build a platform and use the platform. Altogether, these pieces outline a platform for systems engineering modeling and planning on the basis of constructing a suitable platform through various models and exercising the resulting platform, thus improving systems engineering methodology analysis. Specifically, the thesis demonstrates how P-SEMP is the first known technique for SE methodology selection that supports 1) mathematical models of task performance, 2) analysis of SE methods as tasks in a concrete sense, 3) inclusion likewise of soft or subjective criteria, and 4) expandability to investigate new or different SE method proposals in a unified and effective manner.
Showing items related by title, author, creator and subject.
Comparative analysis of the effects of system structure and information input characteristics on the system's response of a military intelligence handling system Maddox, Edward Reeves (Georgia Institute of Technology, 1971-08)
Massey, Kevin C.; Heiges, Michael W.; DiFrancesco, Ben; Ender, Tommer Rafael; Mavris, Dimitri N. (Georgia Institute of Technology, 2006-06)A System-of-Systems design methodology is used to evaluate tradeoffs in the design of a guided bullet system for mortar defense. Guided bullets were designed to match the calibers of four different existing auto guns and ...
Soban, Danielle Suzanne; Mavris, Dimitri N. (Georgia Institute of Technology, 2001-10)The need for a comprehensive framework for the analysis of military system effectiveness is presented. Changes in the world’s economy and its effect on decision making is discussed, as well as the three primary ways decision ...