The definition of complexity is contentious. The lack of a univocal definition for complexity is leading to a cacophony of different views and lexicons, limiting the collective ability to handle complexity effectively. This article aims to resolve this impasse by identifying or developing a useful and acceptable definition of complexity by focusing on how this term is used in practice, and hence determine which definition resonates with the broadest possible community. To understand how the term is being used, several popular complexity-definitions have been identified and de-constructed into component definitional elements. These elements are then compared to the descriptions of complexity in a range of key community documents to infer which definitions most align with the document authors usage of the complex term. The results indicate that new emergent definitions are far more popular than the established dictionary and complexity theory definitions, supporting previous surveys. These usage insights are then analyzed to develop a definition that can unify, as much as possible, the different complexity perspectives. The proposed definition establishes that complex(ity) is when relationships between elements are weaved together such that they are not fully comprehended, leading to insufficient certainty between cause and effect. This definition was tested at a series of international workshops and accepted as the most useful definition.

The INCOSE Complex Systems Working Group Heuristics Team has selected 67 Principles and Heuristics that are considered to be particularly relevant to Complex Systems. These have been incorporated into a Difficulty Assessment Tool that prioritizes the list of Principles and Heuristics based on scoring of a matrix of four Difficulty Elements and six System Elements (to characterize the nature of the complexity). The purpose of this paper is to describe an initial assessment of the effectiveness and usefulness of the Difficulty Assessment Tool. The Tool has been used to assess eight Case Studies by five assessment teams — one with three people working together, one with two people and the remaining three with individual assessments. The results of these assessments have been compared using four different correlation methods, using the total weighted Heuristic score, the maximum weighted Heuristic score, a Match / Mismatch analysis of the top fifteen and bottom seven Heuristics, and a difference ranking between pairs of assessors of all 67 Principles and Heuristics. The last two assessment methods are shown to be more insightful. The assessment teams then reviewed the relevance of the highest and lowest-ranked Principles and Heuristics to the full Case Study definitions (Problem and Outcome). There is good agreement of relevance for the highest-ranking Principles and Heuristics, less so for the lowest-ranking ones. Based on this initial assessment, the DAT shows promise to help people develop complex systems. The paper concludes with recommendations for further assessment of the Difficulty Assessment Tool.

An INCOSE-wide initiative has exposed at least 600 heuristics. Previous work indicates that rationalizing and simplifying this set to make it useful and memorable is difficult, if not intractable. Difficulty Assessment Tools (DATs) have been used for years to characterize the difficulty of a problem and provide tailored advice. This paper explores using a DAT to characterize the problem, and using the outputs to provide heuristic and other forms of advice. To test this approach, 50 heuristics and 10 principles were scored and embedded into an online DAT. An experiment was conducted to determine whether the DAT discussion, recommended approach, and heuristic/principles advice were useful. All teams considered the discussion very useful. As might be expected, the results indicated that the heuristic usefulness was a function of the teams' experience and familiarity with the task. The tool prioritization of suitable heuristics met developers' expectations, but was undetected by the users. This maybe because the heuristics were a hand-picked set of 50 Heuristics from a set of 600+, meaning all were highly useful. Further work is proposed to check this assessment. The DAT usefulness results indicate that Systems Engineers should use the DAT to inform their approach throughout the lifecycle.

Systems of Systems are inherently complex, and hence often, traditional Systems Engineering (SysE) approaches may be inadequate. To assist, the INCOSE Complex Systems Working Group seeks to create and develop a useful set of SysE heuristics that can provide guidance. Analysis of an initial set of heuristics, identified using complex(ity) search terms across an INCOSE database, indicated that they did not sufficiently cover the wider system of interest and culture aspects and required further independent review and usage to become established. This paper addresses these concerns by using additional search terms across the INCOSE database and reporting the findings of an independent SysE team review of the original and newly identified heuristics. Using this approach, an additional 15 heuristics have been added, and modifications to both sets have been identified and agree with an independent focus group. It is concluded that the heuristics identified are useful and have resolved the breadth issues. However, to ensure the heuristics are more useful, additional work is required to rationalize the set from 33 heuristics, explain the rationale for the additional heuristics, and new methods need to be explored to aid the recall of the right heuristic, such as categorisation.

Complex Systems are uniquely challenging by definition and do not readily conform to traditional systems engineering approaches. Complexity Assessment Tools can be used to indicate what elements of a task are complex, but there is paucity of advice to help you act on the data provided. A simple first step to aiding decision-makers with complex challenges is to provide a set of suitable Heuristics. Research conducted by the INCOSE Complex System's Working Group (CSWG) Heuristics team has curated numerous heuristics. This paper reviews the suitability of the developed sets of Heuristics and the applied approach to developing these Heuristics which are then tested against a range of case studies. It concludes that the Heuristics are valuable and useful. However, the approach of seeking a small, pithy, justifiable set of memorable heuristics using bottom-up approaches appears challenging. It considers a range of alternative approaches for ensuring the right advice is available at the right time and recommends exploring the use Complexity Assessment Tools or Complexity Categorisation Frameworks to provide the data to point towards heuristics for the challenge being addressed.

The inability to handle rising complexity effectively is often the cause of project, organizational, enterprise, and even societal collapse. A tractable set of heuristics for handling complexity that can mitigate this risk is consequently highly sought. However, conventional experience-based approaches for identifying complexity handling advice tend to lead to informed but complicated constructs that may be considered over-prescriptive and burdensome for handling complex problems, especially when the need for this support is acute. Further the cacophony of advice, with their tailored lexicons, can cause organizational confusion. This paper explores the development of a simple set of heuristics using an inductive approach that seeks to reduce the decision space and add insight without being overly prescriptive or complicated. An initial set of Heuristics are developed using first-principles. These are then tested and proven by comparison with the dominant discourse in a literature search, to assess if they are simplifying and contributing to established practice, and assessed in a survey to determine if they are useful compared to other similar sets. It is concluded that the proposed set are more useful than similar sets, and that the simplified set of seven heuristics should be developed further to complement other approaches that aim to inform decision-makers in projects, organizations, and society as they seek to handle complexity effectively.

The need to handle complexity has shifted from a nice to have to ensure project and organizational success, to a necessity. One approach to improve organizational performance in handling complexity is to observe what has been successful with similar complex activities in the past. Consequently, suitable Complexity Categorization Frameworks are required. This paper seeks to identify appropriate frameworks by assessing prior art tools & frameworks identified through a literature search and building on these insights to create alternatives. It identifies common failings and suggests that more categories are required. A new framework is created to respond to perceived weaknesses in those reviewed. The usability and accuracy of the reviewed and new frameworks are then tested via a survey. The results indicate that the new and lesser-known frameworks are considered to be better at categorizing complexity, than the more commonly known and accepted frameworks. In addition, the frameworks with a greater number of questions and categories scored highest. To accommodate complexity fully it is recommended that either the established frameworks are adapted to include more dimensions of complexity and categories, or that the approach used by the new evolved question-based framework is adopted.

The Oxford English Dictionary (OED), the established definition of words in the English language, is at odds with other definitions of complexity proffered by Complexity Theory. This variance is likely to cause confusion in the delivery community. The incorrect classification of a project between ‘complicated’ and ‘complex’ is considered by some to be a major source of project failure; implying that resolving this issue is critical to successful system development. This paper explores the definition of complexity by assessing definitions from various sources and by conducting a survey of over 100 delivery professionals. The results demonstrate the extent of the confusion and have informed considerations on how to resolve this. This paper recommends that the definition is either defined at the start, or that the term is avoided by using its component parts. This paper proposes supporting an emerging definition that resolves many of the issues, if adopted widely.

Delivery complexity is recognized universally as continually increasing; suggesting that Complexity or Difficulty Assessment Tools (CATs/DATs) are even more critical for ensuring that the right delivery approaches are selected. However, these tools appear immature, with significant diversity between the tools. Consequently, which tool to use, or type of tool to develop, becomes a critical decision. This paper seeks to identify what a good DAT looks like by extracting and discussing potential benefits from assessing a range of tools/papers and direct observation. It then assesses the three identified categories of DATs - the four-box model, the questionnaire-based approach and the top-down (TD)-based approach - for potential suitability in meeting these benefits. The TD approaches scored well, even accepting the limitation of the assessment. This paper concludes that new DATs should be developed using TD approaches, replacing the questionnaire based approaches, which are difficult to modify, and hence cannot readily keep up with the pace of change.

Complexity, unpredictability, and constraints, such as the need to deliver in short timeframes (speed), can all contribute significantly to how difficult it is to deliver a task. This is especially true if these factors remain unnoticed and hence unmanaged. To manage difficulty it is therefore essential to identify and if possible measure the difficulty within the task. An assessment of this type is valuable to ensure that complexity and difficulty are managed effectively, substantially reducing the risk of failure. The range of difficulty assessments considered either provided a simple difficulty assessment and a limited range of project complexity types (four), with detailed management methodologies for each, or a detailed assessment with limited guidance or mapping to the management of complexity and difficulty within a task. This paper explores an alternative approach to try to obtain the benefits of both approaches. It proposes breaking down difficulty into intricacy, unfamiliarity, unpredictability, and constraints, the former two combining to indicate the complexity. It proposes that these difficulty aspects are considered across the delivery lines of development such as TEPIDOIL or POPIT in a 2D grid. The paper then reviews a difficulty assessment developed using this approach for breadth by comparison with contemporary papers, support to management of complexity and difficulty in a task and through user pilots. The results indicated that the tool developed resolved the issues highlighted in providing a detailed assessment that can inform a range of management decisions.