Complex Systems Design & Management

Cyber-Physical Systems are the next step into globally integrated software systems. They are the result of the combination of embedded system with cyberspace. Cyber-Physical Systems support real world awareness in the Internet and the access to global data and services by embedded system. The engineering challenge for Cyber-Physical System is the combination of characteristic properties of embedded systems such a real time, functional safety, dependability, closedness with characteristic properties of the internet such ad openness, partial availability, restricted quality of service and reduced dependability.

Complex systems resist analysis and require experimenting or simulation. Many enterprise settings, for instance with cases of competition in an open market or “co-opetition” with partners, are complex and difficult to analyze, especially to accurately figure the behaviors of other companies. This paper describes an approach towards modeling a system of actors which is well suited to enterprise strategic simulation. This approach is based upon game theory and machine learning, applied to the behavior of a set of competing actors. Our intent is not to use simulation as forecasting - which is out of reach precisely because of the complexity of these problems - but rather as a tool to develop skills through what is commonly referred as “serious games”, in the tradition of military wargames. Our approach, dubbed GTES, is built upon the combination of three techniques: Monte-Carlo sampling, searching for equilibriums from game theory, and local search meta-heuristics for machine learning. We illustrate this approach with “Systemic Simulation of Smart Grids”, as well as a few examples drawn for the mobile telecommunication industry.

Well-established systems engineering approaches are becoming more inadequate as today’s systems are becoming more complex, more global, more COTS/re-use based and more evolving. Increased level of outsourcing, significant amount of subcontractors, more integration than development, reduced project cycles, ecosystem like collaborative developments, software product lines and global development are some of the changes in the project life cycle approaches. In this paper, we present the results of an exploratory case study which tries to identify the agility problems in large scale software intensive defense projects. This is the first step of our research in which the overall objective is to improve the agility attributes of the traditional systems engineering approach.

This paper is based on the outcome of a research project developed at the RATP Group (the world’s fifth largest public transport company). This ongoing research project aims to provide RATP with a proper methodology that will help it to cope with complexity and to address the challenges of the future market (competition, large programs...). The company has chosen the open method Praxeme and was willing to check whether it could apply to the realm of transportation. Praxeme proposes an efficient architecture framework that links the aspects of the Enterprise System together. This framework lays the groundwork for a comprehensive approach to socio-technical systems; it strongly ties the various models together moving progressively from clarified business knowledge to concrete solutions. A very specific case study was used to demonstrate the value of a modeling approach in the eyes of the decision-makers. The approach proved able to oversee issues in the running of the system.

In 2012, Thales Systèmes Aéroportés initiated a study on the PBS (Product Breakdown Structure) as the central enabler for Systems Engineering throughout the lifecycle of a Solution. The obvious similarities between tree representations suggest in a simplistic and reductive way that breakdown structures are only an unnecessary redundancy due to complementary businesses designating the same reality in different ways. However, a more careful examination of international standards and stepped uses of these tree frames shows how different their purposes are and how they support different features of the system being designed. Nevertheless a confusion exists between these tools. Our paper asks a question of principle: “does this confusion come from imprecise and unclear standards on this subject, from incorrect interpretations of these by system architects or from vague approximations in the use of technical tools that structure and size R&D projects?”.

This paper shows the utility to follow an architecture framework in order to design complex systems with a holistic approach. Multi-objective Optimization techniques extend and complete the architecture framework to support trade-off analysis and decision making in the Systems Engineering design process. The merging and combination of these two approaches, decision making and systems engineering, contribute to the efficient design of systems by helping to meet needs and constraints stemming mainly from the system analysis.

To support this assertion, we present a case study for an Electric Vehicle Powertrain. The decision problem is modeled as a Pareto model, in order to find a solution for the Electric Vehicle Powertrain that maximizes its autonomy and minimizes its total cost of ownership.

The Laser MégaJoule (LMJ) is a laser beam facility dedicated to the inertial confinement fusion program. Its construction started in 2003 at the French Atomic Energy Commission CESTA centre located near Bordeaux. This facility is an extremely complex system, designed using a concurrent engineering process, with state of the art performances. Requirement management is an essential part of the project management of such a complex system, characterized by the integration of multiple heteroclite subsystems and products. This activity ensures the mapping between the system requirements of the project’s end user and the lower-level specifications for subsystems and product suppliers. Using adequate methodology and tools, requirement management acts as a guide throughout the whole integration process, from the first steps of the definition phase to the last part of the project qualification and delivery. As an illustration, this paper presents how the project team (CEA/DAM as the prime contractor and Areva TA as its support) handles alignment performance engineering for the LMJ project using a specific tool based on software package DOORS made by the Telelogic company.

ARCADIA is a system & software architecture engineering method, based on architecture-centric and model-driven engineering activities. It targets systems whose architecture is largely constrained by issues such as performance, safety, security. This paper focuses on functional analysis concerns in ARCADIA, and return of real life experiments around use of existing standards such as Architecture frameworks, SysML/UML, for this purpose.

Managing changes in requirements more effectively than others can become a source of competitive advantage for companies designing and developing complex systems. In this paper, we report that decision-makers in these firms feel that understanding the effects of requirements change in the design process is a difficult task in practice. Such knowledge is however crucial to organize the design process in ways that mitigate for the impact of requirement change. Considering this need, we propose here a modeling framework to determine the impact of requirements change in the design process of complex systems. We view the design process as a requirements-driven process containing steps performed in a mechanical way, but ultimately controlled and steered by the human agents involved in it. Therefore, we develop a technique which allows the modeler to capture both the mechanical relationships and the decision-making behavior of design agents. We introduce attributes such as requirements availability, stability, difficulty and margin, which act as process variables driving the design process. Task properties and the agents’ behavior are then modeled as functions of these process variables, which are dynamically updated as the design process progresses and decisions are taken. The potential of this modeling technique is illustrated on the design process of a turbine blade cooling system. Discrete-event Monte Carlo simulations are used to assess the impact of requirements change during design.

The development of complex systems involves many people from different disciplines, each communicating with his own jargon. These different languages may lead to misunderstandings between stakeholders that cause a significant increase of development costs. This paper addresses this communication gap by proposing the usage of domain-specific tooling, which is shared by all stakeholders. We argue that logistic systems are well suited for the usage of such tooling. This is illustrated by the application of domain-specific tooling in the warehousing domain. We present a warehouse-specific graphical configuration tool built on top of a warehouse-specific language and apply it to an industrial automated case picking warehouse. This application shows that the communication gap between specification and implementation can be reduced by introducing parameterised components and behaviours and local optimisation rules with well-defined interfaces.

When constructing a model of a system or system of systems, one usually decides between utilizing a qualitative model or a quantitative model. There exists a desire to leverage both modeling approaches by finding a way to compare and contrast both types of models. In this paper the authors make attempts to leverage these modeling techniques by exploring the use of a newly formulated methodology, referred to as Relational Oriented Systems Engineering and Technology Tradeoff Analysis (ROSETTA) as a way to compare qualitative assessment and quantitative analysis. This was applied to a sample problem of constructing a model of the Smart Grid. As a proof of concept, instead of examining the entire Smart Grid, only Demand Response and day-ahead load prediction were assessed. Qualitatively, the Quality Function Deployment methodology was used as a representative means to capture Subject Matter Expert (SME) opinion. On the quantitative side, an Agent-Based Modeling approach augmented with elements of Discrete Event Simulation was employed to construct a physics-based model. The use of ROSETTA allows for communication between the SMEs and the modelers, which was used to improve the accuracy of both models. This improvement comes from the iterations of the qualitative assessments and quantitative analyses, successively building off of insights gained from previous iterations. This paper shows the first steps in leveraging the benefits of both qualitative and quantitative models.

This paper describes the results of the first 5 years effort to develop a new focus and process for creating an awareness of systems thinking methodology in pre-college students. The effort was undertaken as part of the University of Texas Pre-Freshman Engineering (PREP) program based in San Antonio, TX. The PREP program conducts a 6 week summer classroom environment on local university campuses for Texas Middle School (7-8-9 grades) students focusing on the basics of Science, Technology, Engineering & Math (STEM.) Our particular effort was to add a fourth year known as PREP IV. PREP IV served as a prototyping laboratory to experiment with new curricula for introducing the systems thinking mindset to pre-college students.

Several approaches, such as Community Based Projects, Student Centered Learning, Activity Based Learning, and Challenge Based Learning were incorporated into the curriculum developed. As a result of the exploratory prototyping fundamental understandings of student basic needs were uncovered including: (1) creating a systems thinking awareness or mindset in the students; (2) developing a teachable approach to grasping the concept of what is a system; (3) learning how to do the abstraction required for modeling a system; and (4) analyzing the behavior of a system from a systems model. Most of these basics of system thinking are not intuitive but require innovative development approaches. This paper discusses teaching approaches that address those new understandings. Specifically, it includes the approach of tightly integrating the study of system thinking concepts with the learning exploration of a local community watershed as a system. System definition, modeling, and behavior analysis in a parallel top down theory and bottom up approach results in system thinking mindset retention that was measured by pre and post testing of students. In addition, the basics of the systems engineering process were built into the curriculum providing a process framework to support the systems concept mindset and awareness.

A high level watershed system model was built using the STELLA computer model. This model was constructed from the ground up in parallel with basic water science teaching of: the water cycle, rainfall rates, rain accumulation, watershed area determination (using U.S. Geological Service or USGS topographic maps and Google Earth), soil analysis, evaporation, and river water flow fundamentals. The model was calibrated using an historical rainfall event along with measured stream water flow rates at local stream USGS monitoring stations. Once calibrated the model was exercised to discover the benefits of reduced flows and flooding associated with creating water retention dams. Soil and surface runoff conditions were varied to develop a student understanding of increases in community land development on the potential for flooding. The model and graphical outputs are discussed in the paper.

As a result of the PREP IV experience the students could clearly articulate an understanding of a very important community system. They showed this in a formal presentation to local senior water system executives, water science university professors, and parents. Continuing PREP IV classes are extending the systematic discovery of how to develop curricula that creates sustainable systems mindset awareness. Areas other than watershed systems, such as aerospace, energy, environment, and medicine are also being introduced into the PREP IV systems curriculum.

The development of embedded systems for complex cyber-physical products involves multiple processes and disciplines – from project management and requirements engineering, to configuration, integration, simulation, test and verification management. Traditional model based systems engineering approaches, where generated models and simulations are largely isolated from one another, make it almost impossible to get a ‘holistic’ view of a complete product or systems behavior.

While most frameworks for model-based design support a single discipline, what is actually needed is a framework that can handle the multi-disciplinary architecture and systems integration of the complete product or system. There is a need for standards to enable the combination of cross-discipline design efforts in a common environment that fully supports modeling, simulation and embedded software generation and validation.

This paper outlines an integrated approach where the Requirement, Functional, Logical, and Physical (RFLP) decomposition of complex cyber-physical products is achieved in a fully integrated 2D / 3D collaborative systems development environment. We will outline how the Modelica¹ language, in conjunction with the open ‘Functional Mockup²’ and AUTOSAR³ concepts, can be leveraged by a Product Lifecycle Management (PLM) based systems integration platform, to define and evaluate the functional definition of a complete product or system. We will illustrate this approach through an example of an automotive climate control system development process.

Systems engineering enables the successful realization of systems, focusing on defining customer needs early in the development cycle. However, when the development of systems needs to rely on legacy designs there is little methodological support. Furthermore, in the automotive domain, product diversity increases system complexity so much, that reuse becomes much more difficult and time consuming. We believe a specific strategy must be adopted to prepare for reuse and to achieve systems engineering by reuse. While product line derivation provides the means to obtain single products form a collection of assets, there is a lack of flexibility and support from a systems perspective. In this paper we present an approach which takes into account systems engineering methodological aspects in product line engineering and provides a means to develop new systems by reusing existing designs. We present the implementation of the tool support for our approach based on the Papyrus SysML modeller and exemplify the concepts through a derivation example of the electric parking brake system.

This paper discusses the concepts of Model-Based Systems Engineering (MBSE) and of Architecture Frameworks (AF) and presents some preliminary results of current initiatives at Renault on these subjects. We advocate the adoption of a MBSE approach, i.e., the application of modeling to support a SE methodology covering the SE design process and activities and supporting the methods that are needed to carry out these activities. This results in the definition of an architectural design framework for the automotive systems development currently implemented in a SysML specialization. It is expected that this work will contribute to foster the reflection on an architecture framework for the automotive industry and stimulate discussions across the automotive community.

This article focuses on bathroom design and more specifically on the sub-sector commonly known as “wellness”. This sector is characterized by its essential element, water. It is possible to generate several types of sensations through water by utilizing several variables such as composition, temperature, and pressure. The leading companies in this sector are allocating more and more resources in the design phase of development. These products need to meet a more demanding market, not only in regards to the product, but also the environment.

This is a growing and complex market because its purpose is not just selling a physical product, but also the types of sensations and emotions that are generated. This set of sensations should result in positive emotions for the user. In short, the aim is to sell ”wellness”. The sensory design plays a key role in this process. It’s also essential to consider the product aesthetics, both the presence aesthetic and the aesthetic in human interaction.

A critical issue in today’s market place is how to obtain the feedback from users. This feedback contributes vital information that is essential in order to improve the features of design for each product. This information can bridge communication barriers between the end user and the product design team: retail, sales department, technical after sales .. each of these units filters the user feedback with a self-interest that lead to the discovery of concrete needs. Therefore providing specific designs that could be vastly improved in terms of interaction and emotional design.

The study of sensory characteristics and user feedback for current products create the basis for future “wellness” designs. The interaction of users with these products and their experience strongly influence research and development. The wellness products are designed to improve the physical and mental health of users through hydrotherapy. By creating more effective sensory and emotional relationships, not only will we achieve greater success in the market, but also enhance the quality of life for the users as well.

Under-frequency load shedding (UFLS) is a technique in which the power grid frequency is used as an indicator to shed down electrical loads at critical periods, which can increase the resilience of the system. We present a multi-scale agent-based model to simulate the impact of UFLS on an energy system. The model allows a representation from a complex system point of view, allowing for crossscale interactions which is not evident in standard grid simulations. Each refrigerator is modelled individually but the population has a heterogeneous configuration. The refrigerators are coupled to a simplified energy system model to represent the grid frequency.Using a simple UFLS strategy we discovered synchronisation effects among the refrigerator population. An emergent oscillation which would be fatal for the system was detected. The model allows to explore the origin of this phenomenon as well as to determine the phase transition which occurs in the system.

Systems engineering is developing everywhere in industry, but human issues incrementally emerge. In particular, this systemic technology-centered approach to engineering rigidifies organizations and lead to surprises. People involved are not fully aware of what is going on. This paper identifies situation awareness and authority issues in complex systems design and management, and discusses possible solutions. It more specifically focuses on an Orchestra model of socio-technical systems.

Building itself is like a complex system with a number of physical processes that interact with each other and with the environment. From the control point of view, this system is composed of multi dynamic subsystems that must be monitored and controlled in order to achieve occupants’ well-being at lowest energy use possible. However, environmental and occupants change in a building, and this increases the complexity in applying control systems. For this reason, this paper presents a framework for the application of systems engineering (SE) concept - a systematic approach for adapting procedures, tools, and standards to all practical problems - in the design of control systems by run-time coupling for building performance applications. Although this is based on SE good practice and corresponding SE standards, the development life cycle of control systems are covered ranging from the operational needs to implementation, operation and disposal. As an essential step toward this goal, namely using computer simulations to evaluate the impact of advanced control systems on buildings indoor operation and energy efficiency, this paper describes a collaborative distributed simulation that was developed and implemented between different simulation tools such as ESP-r and Matlab/Simulink. The paper also addresses a study of verification and validation (V&V) issues within this framework: inconsistency checking, traceability issues and all requirements related in SE standards, especially EIA-632.

This paper presents a five-phase taxonomy of procedures to support the design of complex engineering systems for uncertainty and flexibility. A review of major contributions was done to identify relevant procedures to support initial design generation, uncertainty recognition and modeling, concept generation and identification, design space exploration, and process management and representation. The taxonomy integrates contributions from surveys, articles, and books from the literature on engineering design, manufacturing, product development, and real options analysis obtained from professional e-index search engines. The organizing principles of the taxonomy were developed keeping in mind the intended user: the engineering designer. The taxonomy aims to provide guidance to designers in selecting appropriate tools at relevant design stages for both industry application and engineering education. It also aims to provide a framework to identify and organize ongoing research activities in this emerging research area.