Think of an ant – tiny and rather insignificant on its own. Now think of an ant colony and all of a sudden you have ant path planning, brood sorting and nest climate controls. All this is decentralised; emergent. Now look at man-made systems and organisations; interconnected networks, global connectivity, systems-of-systems thinking. These ideas and implementations are emerging out of typically hierarchical, unresponsive, difficult to maintain legacy approaches grounded in an increasingly fast moving, mass customised and dynamic environment. Given this, the understanding and subsequent exploitation of complex and emergent organisations will have an important role to play in man-made organisations; logistics, manufacturing, co-operation and connectivity. Whilst physicists and to some extent computer scientists have embraced the ideas of emergence and complexity to varying degrees, it will be the engineers and managers that will have to understand and implement them in real life situations. Very little has been undertaken to explore and explain the importance, implications and even benefits of emergent and complex systems in industry. Hence, within the context of the aerospace industry, this thesis is written with the aim of explaining the situation with the engineer and manager thoroughly in mind.
1.1 Background and Motivation
Complex systems often can exhibit behaviour that is not easily predictable when looking at the individual components of the system. The exhibited behaviours are sometimes beneficial and sometimes not. Such behaviour is observable in many natural systems from the swirling patterns of bird flocks to the purposeful social activities of insect colonies. Man-made systems, products and the organisations for their creation and maintenance, can also display such emergent behaviour – often unintended and detrimental. Conversely, benefits can be had when simple systems interact to produce desired complex behaviours where the whole is greater than the sum of its parts.
Explaining this emergent behaviour is a new discipline based on various theoretical approaches including complexity theory, predator-prey behaviour and game theory.
As explained in Chapter 2, BAE SYSTEMS, the sponsoring company envisages a dramatic increase in complexity in three key areas:
- The organisation itself – BAE SYSTEMS is experiencing a generally uniform trend in the products and services it provides – increased technological complexity requiring greater interaction between business units in the company, compounded by the lower lead times demanded by customers;
- The products it manufactures – The defence market and therefore BAE SYSTEMS is moving away from conventional products, placing the emphasis on capability requirements (e.g. anyplace, any environment, within a day) and information warfare instead of product specifications. This is leading to an emphasis on non-conventional product solutions that may potentially be autonomous, cheap and numerous instead of manned, expensive and non-abundant;
- The logistics, manufacturing and support structure – As large organisations in the aerospace sector divest of non-core business and focus on the design and support of products, a large network of third party suppliers has emerged. This network is being increasingly integrated to companies such as BAE SYSTEMS through extranets, allowing transparency of information. Cross organisational collaboration is also on the increase. Furthermore, smart manufacturing and purchasing systems are being integrated into the design process.
These areas share the following commonalities; numerous members (I refer to them generically as agents from now on), high levels of interaction, common global and conflicting individual goals, all potentially leading to emergent behaviours. The company is therefore aggressively assessing new and novel interaction paradigms such as self-organisation, conflict resolution and “control without controlling”. These concepts can be rolled up into the study of Multi-Agent Systems, or MAS. For our initial purposes MAS is defined as a collection of autonomous members socially interacting explicitly or implicitly.
The objectives are two-phased. The scope of this research must inherently start off by looking at potential applications and state of the art research and then focus in on a viable commercial application. The first phase of the objectives is therefore to:
- Examine the trends in the aerospace industry and BAE SYSTEMS in particular. Assess the commercial value, trends, implications and application of the Emergence/MAS body of knowledge to BAE SYSTEMS’ current and future operation.
The second phase of the main objective leads on from the initial phase. This is to:
- Explore in detail the most promising potential application or body of knowledge that may provide BAE SYSTEMS with increased knowledge, and thus a competitive advantage.
The final objectives are to:
- Bring the body of knowledge in MAS and similar disciplines to a level that is accessible and practical for engineers and managers to consider. This is predominantly through publications and a thesis written with such an audience in mind;
- Produce and defend a thesis that contains a contribution to knowledge of the potential for practical applications of emergence/MAS and how this relates to trends in the aerospace industry;
- The validation of an exemplar application, either theoretical or practical, of MAS, which may have commercial or intellectual value for BAE SYSTEMS;
- Attain, or be very close to attaining the Chartered Engineer (CEng) status by the end of theEngD.
1.3 Research and Thesis Structure
In order to explore the open ended issue of complexity and emergence in the aerospace industry the research has been broken down into seven main areas:
- Informal definition of objectives;
- Detailed examination of industrial trends and requirements;
- Detailed examination of trends, concepts, commercial application and issues of industrial and commercial research related to (2).
- An examination of the fit between the requirements in (2) and findings in (3). This is known as a gap analysis. Key “gap” issues identified in this analysis are then the focus for the subsequent steps;
- Firmer definition of objectives;
- Detailed examination of promising leads;
- Discussion of future work.
This process is illustrated diagrammatically in Figure 1.3:1.
FIGURE 1.3:1 – DIAGRAMATIC ILLUSTRATION OF RESEARCH PROCESS
This thesis is structured around the implemented research methodology and the chapter descriptions are provided below.
In Chapter 2, the aerospace industry is examined in detail. The direction of the industry and BAE SYSTEMS in particular is considered. High level issues surrounding emergence and complexity are highlighted. Chapter 3 then explores the role of BAE SYSTEMS’ ATC Sowerby – the location of my placement – with regards to how it aims to fulfil the commercial requirements of the company. At this point a closer look at emergence and complexity is provided. Chapter 4 then takes a detailed look at the theoretical aspects behind complexity and emergence and issues of control and autonomy that are closely tied to emergence. In Section 4.7 a detailed flocking model created for the purposes of examining emergence is described and analysed.
Chapter 5 then looks at Multi-Agent Systems (MAS) which were identified as an important and useful way of exploring emergence. MAS are explored within a theoretical as well as a commercial context. Key arguments about the state of MAS research with regards to emergence and complexity in a commercial setting are outlined. Chapter 6 then looks back at the key issues raised in the previous chapters, examining research focus versus industry trends and solidifies the initial research brief outlined in this chapter. A crucial gap identified is the lack of understanding and formalised research around the relationship between organisational structure and organisational performance.
Chapter 7 formalises a set of generic organisational structural metrics, while Chapter 8 describes a bespoke Java based simulation test bed, along with specific performance metrics, to examine this relationship. Chapter 9 then explores the results and details the data mining and data reduction methods used to examine this complex relationship. Chapter 10 then discusses the implications of the findings in the gap analysis and the subsequent results. Furthermore, a look at potential commercial applications and future research is covered.