A methodology for identifying flexible design opportunities in large‐scale systems

Despite many uncertainties, industrial fields such as energy (eg, oil rigs), utilities infrastructure (eg, telecommunications), space systems, transportation systems, construction, and manufacturing make large and mostly irreversible investments in systems with potentially long life cycles. The life cycle value of such systems may be increased significantly by designing in the flexibility to make future changes. This paper presents a systematic methodology for designers to identify flexible design opportunities (FDOs) more comprehensively and earlier in the system design process. The methodology guides designers to generate flexible design concepts that can be assessed, selected, and integrated into a design during an early stage of the design process. We demonstrate and validate the FDO methodology on a case in the offshore drilling industry.

dynamic strategy for value sustainment. 9 According to de Neufville and Scholtes, 5(p.39) accounting for flexibility early, during the design phase, can increase financial performance by at least 25% over standard design procedures (even when accounting for the costs [eg, extra design effort and potential performance reductions] that flexibility sometimes adds). A flexible system enables the exercise of options for smaller, more frequent upgrades that extend the system's useful life. 2 This paper targets the potential of increasing life cycle value by designing systems for flexibility, early in the design process, during need identification and conceptual design, to embed flexible design opportunities (FDOs) that enable its users to deal with unfolding uncertainty during system operations. Embedded flexibility provides the most value for industries with expensive, unique, enduring systems that face large uncertainty in user needs. 10,11 This paper addresses the need for flexible design methods, organized into a consolidated framework, 1 with an extended perspective of the technical system that includes external interactions-ie, an "endto-end" methodology 12 that guides the user from change initiation to flexible design concepts (FDCs) and solutions. We focus on the concept generation phase, where "more research is needed to develop new procedures or adapt existing ones." 13 We derived the method based on the need in the upstream oil and gas industry, where-due to social, technical, and environmental complexity, and the limits in planning operations strategically during design-a strong lack of uncertainty handling was observed, leading to design changes throughout the life cycle and making upgrades very expensive and suboptimal, especially postdeployment. 14 Hence, this paper presents a systematic methodology for designers to identify FDOs more comprehensively and earlier in the system design process. The methodology guides designers to generate FDCs that can be assessed, selected, and integrated into a design during an early stage of design. Due to space constraints, this paper presents only an overview of the FDO methodology; many further details are available in the first author's dissertation. 15

CONTEXT AND RELATED WORK
We organize our review of related work around Cardin's 1 five-phase framework for enabling and managing flexibility in engineering systems ibility may be generated, evaluated, and deployed. As the arrows in Figure 1 indicate, designers do not have to follow this process sequentially; rather, they may go back and forth between phases or explore the phases in any suitable order.
In addition to Cardin, 1 several others have explored "end-to-end" methodologies for identifying FDOs by guiding a user from exogenous Change Drivers (CDs) toward FDCs and solutions. (The Appendix contains a list and definitions of the italicized terms used throughout this paper.) These methodologies generally emanate from the fields of "engineering systems" and "manufacturing and factory planning" (Table 1). Moreover, several matrix-based methodologies (introduced in the next subsection) include the literature from the field of product development. The methodologies of engineering systems often target specific application contexts (eg, construction), although they may apply across various other domains as well. Based on the definition by Chmarra et al, 16 they all relate to "lifetime adaptability" (prolonging the product's service life in its normal operational mode and adapting it to new operational modes) and also partially to "runtime adaptability" (increasing utility through flexibility when the product performs a task, such as the ease of switching between two tasks). Some of the methods, eg, Refs. 17 and 18 target operational flexibility by easing the configurational changes of "systems of systems." The FDO methodologies in Table 1 led us to several observations regarding the five phases in Figure 1: • All of the FDO methodologies in Table 1 focus mainly on phases 2-4 in Figure 1. • The FDO methodologies begin variously in phases 1-3, although these phases are not performed in a strict, consecutive order (as allowed by Cardin 1 ).
• All FDO methodologies in Table 1 aim to specify FDCs. Most also yield evaluated solutions that provide a starting point for embodiment design.
• Although the FDO methodologies in Table 1  • Some FDO methodologies 17,31 intentionally embed the selection of flexibility-relevant aspects (eg, selecting changeability type, suitable evaluation methodology) as an output of a separate step, while others constrain such aspects and pursue a specific goal from the beginning (eg, targeting "modularity" of Change Objects only 26 ).
These observations point to the need for a more comprehensive, "end-to-end" FDO methodology that better spans and aligns with Figure 1. Furthermore, comparing the available FDO methodologies against the needs of industry (mentioned in Section 1)-including support for system suppliers in early design phases to embed flexibility successfully into technical systems-also reveals many shortcomings and motivates the need for further work in this area. In this paper, we take up this challenge and present an overview of such a FDO methodology.
As a key aspect of several of the available FDO methodologies, we also noted that many have employed matrix-based methods (Table 2) such as the design structure matrix (DSM) 34(p. 9) to identify effective FDOs more systematically. Examples include: Mikaelian et al, 27 who sought to identify real options to enable operational flexibility for systems-of-systems (SoS); Koh et al, 28 who targeted "change options" in developing products; and Kalligeros, 23 who focused on customized components in developing products. Engel and Browning 35 and Engel et al 36 focused on modularity as a solution for dealing with future change by comparing alternative allocations of components to modules-in particular, merging or splitting modules of the initial baseline design-to find a product architecture better suited to future changes. The scope of these matrix-based methodologies varies as highlighted in Table 2-wherein the last column compares with the "end-to-end" FDO methodology presented in this paper.
All of the methodologies in Table 2 assume an initial, baseline design as an input for FDO identification-with the exception of Hu and Cardin, 33 who included identifying the best-performing baseline design before considering flexibility. Although most of the methodologies in Table 2 touch on uncertainty recognition and scenarios, only a couple embed them as core elements. With the exception of Mikaelian et al, 27 most of the approaches in Table 2 focus on the identification of Change Objects, direct change propagations among them (addressing indirect ones only partially), and their subsequent prioritization and selection-where "change propagations" entail knockon effects to other objects and are especially important to consider in complex products with strong interconnections. 37 Meanwhile, the TA B L E 2 Comparing matrix-based methods for identifying FDOs Transitions ("types") and CEs ("mechanisms"). However, their multidomain matrix (MDM) approach largely bypasses the identification of Change Objects and does not consider change propagation. Most FDC evaluations use "real option valuations" for large-scale systems as a means of quantifying the value of FDCs when performance parameters cannot be clearly assigned to monetary payback functions. 39(p.6) Of all of the methodologies in Table 2, Hu and Cardin 33 come closest to the more comprehensive, end-to-end (as described in Section 1), matrix-based FDO methodology presented in this paper. Although all of the other approaches in Table 2 are partially matrix-based, none of them maintains a matrix-based approach "end-to-end." Furthermore, they focus on the processing of empirically derived data, which are then used for analysis in the absence of domain experts, but this ignores the available, tacit knowledge of domain experts and limits applicability across varied industrial settings and projects.
The FDO methodology presented in this paper addresses these gaps by covering all of the early design phases-from baseline design and the sources of uncertainty to concept generation and the evaluation and decision on FDCs-with a more holistic, end-to-end per-

PROPOSED FDO METHODOLOGY
Our proposed FDO methodology consists of three interrelated models: an eight-step procedural model, a data model, and an execution model.
The following subsections describe each in turn. We realize that these descriptions involve many specialized terms (which, to note again, are summarized in the Appendix). Section 4 applies the FDO methodology to an example.

Overview of the eight-step FDO Procedural Model (FDO-PM)
Drawing upon Figure 1, the procedural model identifies FDOs through three main stages containing eight subordinate steps, as shown in FDCs and integrating them into the initial reference design. In accordance with de Neufville and Scholtes, 5(p.99-127) this third stage emphasizes the efficient short-listing and selection of candidate designs.

F I G U R E 2
Overview of the FDO-PM with three main stages and eight subordinate steps Consistent selections within and across these three stages are pursued to ensure satisfactory results, which is why these stages' subordinate steps are represented in Figure 2 as layers connected in series or parallel. The following subsections present each of the eight steps in greater detail, with full details available elsewhere. 15 Section 4 follows with an example implementation.

3.1.1
Step 1: Select and adapt the most suitable reference design The reference design must first be proven to fulfill the main requirements. Because requirements for large-scale systems usually vary, and efforts for complete redesigns are impractical, inefficient, and at risk of introducing system malfunctions, the starting point for step 1 is to determine a reference design that best meets the requirements for the project under consideration. Similar designs are usually created by modifying the existing designs based on varying requirements. 40 Hence, after (i) determining a suitable reference system and layout, the reference design is modified where requirements are violated, which includes the modification of (ii) systemspecific items such as products and bulk items and (iii) the system configuration or layout. These activities result in a tentative solution, consisting of a set of Objects that may already include some flexible Objects (as the initial reference design usually accounts for some requirements concerning future changes from the previous project). For instance, some machinery such as cranes might already be overdimensioned for dealing with larger functional or environmental loads than currently necessary. Hence, in step 4, those Objects might not become Change Objects because they already handle future changes without requiring additional CEs (eg, over-dimensioning, modularity).

3.1.2
Step 2: Define Baseline Objects  Figure 2), designers assess the general need for making those Objects flexible and remove irrelevant Objects, yielding a set of Baseline Objects relevant to flexibility.

Step 3: Recognize change sources
This step involves identifying the "change sources," the underlying factors potentially causing Objects to change. Change sources include both exogenous CDs and endogenous System Requirements (SRs). First, the set of CDs (or "system drivers" 12(p.73) ) represents both root causes and resulting ones that lead to nonfulfillment of SRs, and thus, system change. They "represent uncertainty sources that are known to engineers to have significant impact on anticipated life cycle performance" 38 but can also include knowable unknown-unknownsie, "knowable unk-unks." 41 They come from outside the system and will be resolved in the future-eg, oil price fluctuations, health safety environment (HSE) requirement changes, etc. Second, the set of SRs defines the capabilities the system must demonstrate. Meeting a SR provides both value and utility to the system's customer or user.
CDs may be modeled as a cause-and-effect network of uncertain factors, whose changes can affect particular SRs, which, in turn, can affect Iterations may be required (path B in Figure 2) when a scenario identifies new SRs that were unaccounted for in the reference design from Step 1. Objects of dedicated capability would have been omitted from the reference design as future scenarios were not yet accounted for, and consequently, were also not part of the imported Baseline Objects.
If reference designs from previous projects cannot be found with such a capability, "Object placeholders" must be added to the existing Baseline Objects (in Step 2) to ensure that flexibility is considered for those

Objects.
At the next stage, it must be determined if the defined Baseline Objects (Step 2) are significantly affected by the changing SRs.

Step 4: Screen for Change Objects
The Baseline Objects now confront the affected SRs (path C in Figure 2 Figure 2) in order to maintain focus going forward.
Considering only direct but not indirect relationships among elements ignores valuable opportunities to embed flexibility. 33 Hence, changes may include "planned changes" triggered by exogenous uncertainties as well as those induced by change propagation. 29 When Objects are coupled, especially physically, a change made to one typically requires changing the other. 42 Thus, Change Objects may be affected directly by SRs changes and indirectly by physical or logical changes propagating across Objects. Change propagation depends on the selected Transition (Step 5). For instance, changing an entire Object is more likely to have physical implications on connected Objects than merely replacing subassemblies of that Object. Consequently, an iteration, whose scope strongly depends on the product architecture of the reference design, is required between the first and second stages to identify indirectly affected Objects (path F in Figure 2) based on the selected Transition. The identified Change Objects are now transferred to Stage II (path E in Figure 2) to generate FDCs.

Steps 5 and 6: Identify suitable Transitions and Change Enablers
With the set of Change Objects identified, designers can now select rele- which might also require other specific CEs as a prerequisite-ie, they cannot be embedded without other CEs in place.

Transitions and CEs both represent "design variables" to generate
FDCs for the previously defined Change Objects. In this regard, two alternative approaches for FDC generation apply, depending on the project's boundary conditions. In the first, a Transition-driven approach, the CEs are selected (Step 6) based on previously identified Transitions for each Change Object that is frozen in Step 5. By feeding the FDCs to Stage III, they are assessed (Step 7) and provide the basis for decision-making (path K in Figure 2). The best performing FDCs are then determined (Step 8), which represent or can be combined into Flexible Design Solutions. This Transition-driven approach is recommended when upgrade preferences are well established, upgrades have a fixed schedule, and/or the project faces strong time pressure.
In the second, Enabler-driven approach, Transitions (Step 5) and CEs (Step 6) are regarded simultaneously, resulting in a highly iterative process (paths G and H in Figure 2). This allows the identification of best-performing combinations of Transitions and CEs for the identified Change Object. Here, Transitions are considered to be design variables that may be reset and can vary for a specific Change Object in the search for the best CE-Transition combinations. However, to attain a single Transition, which is required for the consideration of change propagation (path F in Figure 2), designers must perform an intermediate assessment (Step 7) to decide on the best Transition(s) for the Change Object with respect to the interdependent CEs. In contrast to the first, Transition-driven approach, where the Transition is defined at the very beginning, the Enabler-driven approach requires returning to Stage II postassessment (path J in Figure 2), where, based on the assessment results, the most suitable Transition(s) are then selected for the Change Object (Step 5). This is the basis for identifying downstream Objects affected by change propagation as they depend upon the Transition (path F in Figure 2). Those indirectly affected Change Objects, in turn, would then be subject to the same, iterative process.
The Enabler-driven approach is recommended when high-performing solutions should be embedded and/or the project faces less cost and schedule pressure.

7: Assess Flexible Design Concepts
When applying the Transition-driven approach in Step 6, the assess-

Step 8: Select Flexible Design Solutions
Based on the assessment results, designers filter and integrate highperformance FDCs to arrive at Flexible Design Solutions, which they may then integrate into the reference design (path L in Figure 2).
However, further iterations of the FDO-PM may still be useful. 43(p.32) As each Change Object might still have insufficient solutions, initially neglected Baseline Objects may be reconsidered by lowering the critical risk threshold (path M in Figure 2). Less critical Change Objects could also prove worthwhile for embedding flexibility, as they might still provide positive value across the life cycle.

FDO Data Model (FDO-DM)
The Transitions are externally imposed change strategies for physical Change Objects to make them refulfill any violated SRs. We define a set of eight Transition operators (Table 3)  An exception is the CEs representing heuristics, which can vary in their level of abstraction from general design principles to very specific design guidelines, [47][48][49] and thus, are applicable for both variant design (changes within defined parameter ranges, such as "add crossbracings to a derrick structure") and adaptive design (changes involving new parameters or new parameter ranges, such as "provide space around Object for spatial expansion"), as suggested by Koh. 46 This also applies to Enabler Reference Objects, which can be a specific Object or a generically defined part of the system (eg, structure or room). This leaves users with several degrees of freedom in applying CEs or Enabler Reference Objects in the application context while ensuring maximum reuse of project-independent knowledge.
Whereas the procedural aspects (FDO-PM) are application-contextindependent, the data stored within the domains (elements, relations) are not. Nevertheless, insights, especially Transitions and CEs, can be said to apply across different but similar contexts (eg, across plant engineering). In general, the extension of elements and relations of domains (eg, CDs and CEs) in the FDO-DM is a continuous process for the application context of concern to increase the value of the FDO methodology once it is applied in projects.
The next subsection provides an example of the DSMs and DMMs in stages I and II of the FDO-PM as they develop through the FDO-EM.

FDO Execution Model (FDO-EM)
The FDO-EM applies the data stored in the FDO-DM to a specific project by running through the FDO-PM, as exemplified in

FDO-EM scaling
The FDO-EM is scalable to four cumulative paths of increasing scope and fidelity, depending on a project's circumstances. The higher paths build on the lower ones, including their results (eg, the Advanced path builds on the Light one). Overall, the FDO-EM is designed to scale flexibly to accommodate a variety of real-world conditions and contexts such as these. The four paths represent a set of alternative levels as recommendations.

Tracing and scenario building
The Here, we distinguish forward and backward tracing: Those scenarios must account not only for the CDs but also for the direction and magnitude of changes those elements will face, as shown by the anticipated future characteristics (eg, increase of drilling depth to 30 000 ft). In contrast to accounting for alternative developments of CDs (eg, 10 000 f. versus 30 000 ft drilling depth), it targets "focused planning" 19(p.101) with only one future development (eg, drilling depth up to 30 000 ft) per scenario. Hence, multiple scenarios are considered, not because of alternative developments of CDs, but to account for independent, causal networks of FDO elements. Hence, scenario building targets a comprehensive representation of consistent or independent future scenarios for which the customer has interest to prepare.
Ultimately, each scenario should have CDs that can be directly related to the SR domain, which is the basis for performing interdomain tracing.
The FDO methodology considers this assessment and decisionmaking in two ways: On the one hand, they are already addressed in stages I and II by assessing their relevancy on an element-by-element basis, as some potential relationships might not be applicable or relevant in a particular context and hence require no further consideration. The identification of Change Objects (Stage I) depends on certain criteria that have already been highlighted by various authorseg, Refs. 12, 24, and 33. In general, they concern both the probability and (switching) cost of changes. Through Tracing in M 1 , the poten-tial relationships among CDs are checked by considering their probability of occurrence (P CD ). The probability of the instigating element (eg, P CD_a with "a" indicating the instigating CD)-ie, the presumed root cause-is put on the element itself. The other probabilities in matrix CD-CD (M 1 )-eg, P CD_ab with "b" being the CD affected by CD "a"-concern the relationships among CDs-ie, the likelihood of impact between CDs once a change occurs in the instigating CD.

AN EXAMPLE APPLICATION OF THE FDO METHODOLOGY
In this section, we demonstrate an application of the Advanced path of the FDO methodology to the design of an offshore drilling platform-a real project that helped develop and implement the FDO methodology on the side. (Further details are available elsewhere. 15

Use case basis
In conventional offshore drilling, performed under atmospheric pressure, drilling fluid and gas exit the top of the wellbore through a diverter and flowline to a mud gas separator (MGS), which separates the gas from the liquid. Solid control equipment such as a shale shaker separates the fluid from solid rocks, and after treatment, the purified mud can be reinjected into the wellbore. In a stable state of drilling with circulating mud, the bottom-hole pressure P BH in the wellbore equals the hydraulic pressure (P Hyd ) plus the dynamic pressure due to annular friction (P AF ). When the circulation ceases, due to, eg, making up drill pipes on the rig, P BH exceeds P Hyd , which can cause an influx that must be circulated out of the well. Such well control incidents come in addition to the repeating "kick-stuck-kick-stuck" scenarios triggered by lost circulation, stuck pipe, and wellbore instability. All of these contribute to nonproductive time (NPT) on the rig. 51 In contrast, MPD utilizes a closed vessel, where fluids exiting the borehole are not exposed to the atmosphere. In addition to the static, hydraulic pressure (P Hyd ) and the dynamic pressure (P AF ), a back pressure (P Back ) can be applied flexibly to the closed system, whereby

Initial situation and preparation
The use case is based upon a long-term tender, in an operatordominated system development, for a drillship that is run collaboratively between a drilling contractor and an oil company. The project addresses the need for drilling operations in increasingly tougher conditions as fields mature and more challenging operating environments are selected. Those environments are usually in deep water (up to 10 000 ft) and drilling depth (up to 40 000 ft), with the rig and drilling systems being exposed to extreme reservoir pressures and temperatures. The deep water also increases the seawater overburden, leading to narrow pressure windows that complicate conventional drilling (due to frequent "kick-stuck-kick-stuck" scenarios) and cause a negative impact on NPT. The objective of the rig is to foster efficient and safe drilling operations, including well completion and intervention.
In this use case, (a) the main technical specifications of the hull design are already set, and system suppliers are now competing in a tender to provide suitable drilling systems to meet those specifications. Based on the two new CDs "j" and "m," we perform interdomain Tracing to identify further, as-yet-unarticulated SRs (Step 5 in the FDO-EM). The confirmation of potential relationships is based on a high P SR and customer-specific acceptability thresholds. In this reduced matrix M 2 *, CD "j" has no impact on other SRs, whereas CD "m" affects SR 1.
Consequently, the core scenario consists of the governing CDs "h," "j," and "m."

Running the Advanced path in the FDO-EM
Building upon the Light path, the Advanced path identifies indirectly affected SRs (Step 3 in the FDO-PM). SR 1, "Active Heave Compensation [Capability]," was not directly articulated but rather identified by Back-tracing and Tracing. It and SR 5, the customer-articulated "Managed Pressure Drilling [Capability]," provide the starting point for further Tracing, as noted by Step 6 in Figure 7. The confirmation of potential relationships depends on high P SR and customer-specific acceptability thresholds. While SR 5 has potential impacts on five other SRs, SR 1 does not affect any others. Embedding the "MPD capability" will certainly affect SRs 2 and 3, as new equipment must be installed and powered. SRs 6 and 7 may be affected as well, as increased capacity is required for the additional gas that is returning from the formation due to operations close to pore pressure, while increasing the mud flow and trapped gas compared to conventional drilling. SR 3, in turn, itself requires increased electric power capacity (SR 2): This allocation requires intradomain Back-tracing (step 7 in Figure 7). Due to the previous allocation by Tracing, however, SR 2 has already been selected.
Hence, in this case, intradomain Back-tracing leads to a double allocation of SR 2.
Based on the indirectly affected SRs, designers can now trace Objects

Assessing FDOs with the FDO-EM report
Once specified, FDCs may be assessed (stage III in the FDO-PM).
However, selections performed within the FDO-EM are not especially amenable to assessment in that format. Thus, we export the FDCs into a formatted table called the FDO-EM Report, exemplified in Table 4 for a derived FDC of the MGS (with the "Adding" Transition and CE 4). "oversize Object with regard to stress/load cases").

F I G U R E 8 Generating Flexible Design Concepts (FDCs) in the FDO-EM
The assessment process refers to the four steps introduced in Progressing to the third step of the assessment, designers benchmark the FDCs relatively using performance criteria. We weight each criterion in the three main performance categories individually and progressively. We then rate each CE for the applicable Change Object (here: MGS) and Transition (here: "Adding" or "Passive") by comparing it to the Baseline Object, multiplying by the weighting, and then normalizing. We add these results to the performance profile shown in Figure 9. Hence, if different FDCs exist, each of them would usually have a different profile, performance rating, and ranking. The overall performance rating is derived by summing up the performance ratings in each category (eg, upgrade effort) for each FDC. Although the performance ratings and rankings support decision-making, a comparison of the full performance profiles of various FDCs should also be considered, as these F I G U R E 9 Performance profile of FDCs for "mud gas separator" profiles may illuminate more nuances than the mere performance rankings.
In Figure 9, to determine the most suitable Transition for the FDCs of the MGS, we compare the five CEs for "Adding" to the one for "Passive" (CE 20), which is in both the left (upgrade effort) and right (upfront effort) sides of the chart. The overdimensioned MGS (CE 20) also has significant operational deficits compared to a fit-for-purpose, rigid design (the other CEs), especially due to the higher operational and maintenance costs. This observation is reflected in the only positive (bad) "operational effort and losses" rating (0.78), resulting in the weakest performance among FDCs and the sixth-place ranking for CE 20. Its rating also depends on the usage frequency of the overdimensioned MGS (influencing criteria considered in the fourth step of qualitative assessment), which will be "once" or "never" (a single anticipated upgrade across the life cycle). Here, the customers of that project also represent the system users who directly benefit from lower life cycle costs, so the FDC with CE 20 is deemed suboptimal to ones that consider "Adding" a MGS. Hence, we remove the solution with the "Passive" Transition and henceforth focus on the "Adding" Transition for the Change Object MGS. Next, we may further assess the other FDCs to generate the best-performing Flexible Design Solution(s) (Step 8 in the

FDO-PM).
All remaining CEs for "Adding" a MGS have beneficial performance ratings and a high frequency of utilization, as they enable maintenance, repair, and overhaul activities. In particular, CE 3 has a very high utilization rate, as the hatch in the ceiling for removal/installation of the Object can also be used by other Objects for multiple purposes. Two CEs

Reflections on the example application
After the example application, to provide further insights on the utility and usability of the FDO methodology, the first author interviewed a dozen workers involved from the drilling system supplier.

CONCLUSION
The FDO methodology presented in this paper is divided into three The integration of further features discussed in literature, such as compound options (ie, an "option on an option," such as combining "Relocation" and "Partial Replacement"), 18,54 which had no applicable use case in this research, could be a subject of future work. The scenario-building process could be extended to allow one to shift from the current "focused strategy," with one possible future, to a "future robust" strategy, accounting for varied attribute values of CDs.
Concerning implementation in a software tool, a stepwise imple- allow FDCs to be defined in greater detail. In addition to extending aspects already incorporated into the FDO methodology, future studies could also consider the "process management" of Concept Generation (Figure 1), "a rich environment for novel research contributions." 13 Closing the gap between the changes envisioned during design and the ones exercised during system operation is also of great interest, as it these aspects, the value of the FDO methodology will increase, as will the ability to augment the conceptual design process with artificially intelligent design tools that account for many future uncertainties in a system's operational use. The best-performing FDCs that already represent solutions or can be combined with existing ones

Must-change Enabler
CEs that must be in place as a prerequisite for enabling the exercise of specific Transitions Object (OB) Physical constituents of a system that can be influenced by a stakeholder (eg, a system supplier that can affect design of Objects)

Prerequisite Change Enabler
CEs that must be in place as a prerequisite for the exercise of other necessary CEs System Requirement (SR) A statement of system capability that is: qualified by measurable conditions, bounded by constraints, and can be validated Tracing Forward-oriented identification of FDOs based on an instigating element Transition (TR) Externally imposed change strategies for physical Change Objects to make them refulfill any violated SRs