Go to the table of contents.

Previous ...


4.3 Systems theory and systems engineering

4.3.1 Systems engineering and systems

Systems engineering (SE) as a discipline in general is focused more on what in the product design. This can be compared to more traditional European design methodology that has a deeper focus on the how of the design task. Systems engineering, coming from a systems oriented perspective, views the design as a system. This promotes a stronger focus on issues like the system boundary, system interfaces, and the collaboration of the components (system elements) of the system. The systems view of systems engineering also leads to a stronger focus (compared to traditional European design methodology) on the verification and validation of the complex collaboration between the components of the system that is necessary for the system to deliver its required functionality. The purpose of this thesis, however, is not to elaborate on the similarities and differences between systems engineering and traditional European design methodology. Systems engineering is introduced in this thesis because the concept of configurable components (see section 5.4) is largely based on the systems oriented view of design. The iterative approach to development used within systems engineering is shown in Figure 24.

Figure 24: The systems engineering process (SE Management Guide, 1990).

The roots of systems engineering can be traced back a long time. However, the initial formalization of the systems engineering process for military development began to surface in the mid-1950s in the ballistic missile programs in USA. In these programs engineering discipline specialists emerged. Each of these specialties not only needs to take data from the overall process, but also to supply data, in the form of requirements and analysis results, to the process. In 1969 the MIL-STD-499 was issued to assist in defining the systems engineering effort in support of US defense acquisition programs. This standard was updated in 1974 to MIL-STD-499A. Another substantial reference to systems engineering is provided by Blanchard & Fabrycky (1981). Systems engineering is both a technical process and a management process. Systems engineering has several similar definitions that has been provided by different sources. The definition given here is from the Systems Engineering Management Guide (1990):

"Systems engineering is the management function which controls the total system development effort for the purpose of achieving an optimum balance of all system elements. It is a process which transforms an operational need into a description of system parameters to optimize the overall system effectiveness".

There are many possible definitions of a system. A concrete and useful working definition is: "A system is a purposeful collection of interrelated components that work together to achieve some objective" (Sommerville, 2001). Roozenburg & Eekels (1995) state that the system is a totality; each element of the system is only what it is, in and through the system of which it is a part. The system is transforming; it undergoes - whether or not cyclic - an infinite series of states. Despite the transformations the system remains within the boundaries set by the structure. That means: the system always transforms into a state that again shows the structure of the system (if this is no longer the case, an explosion in a machine, for example, then we rightly speak of destruction, but then the object is no longer the original system). The structure therefore remains intact through all transformations. Systems are often hierarchical in that they include other systems. A characteristic of sub-systems is that they can operate as independent systems in their own right. Their behavior in a particular system, however, depends on its relationships with other sub-systems. The complex relationships between the components in a system mean that the system is more than simply the sum of its parts. It has properties that are properties of the system as a whole. These emergent properties (Checkland, 1981) cannot be attributed to any specific part of the system. Rather they emerge only when the system as a whole is considered. Examples of emergent properties are: the overall mass of the system, the reliability of the system, and the usability of a system. The overall mass of the system is an example of an emergent property that can be computed from individual component properties. The reliability of the system depends on the reliability of the components and the relationships between the components. The usability of the system involves not only the system itself, but also the users of the system.

4.3.2 Open systems (system theory)

The concept of open systems emerged in the field of systems theory in response to many of the shortcomings of the mechanistic viewpoint (Hitchins, 2003). In modern biology, the open system is fundamental. The human body is an open system, as are many of its internal organic sub-systems. Human activity systems, organizations, and many technological systems are open systems. Some systems may be considered more open than others, but all systems must be open to some degree; otherwise we would not be aware of their existence.

Figure 25: Systems recursion model (Hitchins, 2003)

The basis of the open system model is the dynamic interactions of its components. The open system ingests, and removes waste. The open system responds to stimuli. The open system can exhibit growth, can be stable at high energy levels, and can collapse and disintegrate. The recursion model (Figure 25) presents several interacting systems. Each system is receiving energy and dissipating energy. Each system is receiving resources and discarding residues. Each system is receiving information and sending information. The systems are interconnected such that the dissipation of some systems provides the energy input to others, and the residues from some systems form the resources of others. This is an organismic view of systems. For both energy and resources there is a sense of conservation. For each system (sphere in the model), energy supplied will equal work done, plus energy stored, plus energy dissipated. Similarly, resources supplied will equal resources utilized plus resources stored plus residues. Information is different, as indicated by the dotted line. Information is not a conserved item; it can be given away without being lost.

In Figure 25, the three systems (spheres) each receive information, energy, and resource, and each system also gives out residue, dissipation, and information. The grey envelope is also a system, but at one hierarchy level higher. Note that it also receives information, energy, and resource, and gives out residue, dissipation, and information. This is the basis of recursion and encapsulation (see section 4.3.5). If enough of the right sub-systems were brought together, such that each received what it needed from the others, while each gave out what the others needed, then there would be a complementary set of systems, which might form a stable, higher level system. If such a system could be could be described/labeled adequately, then in some contexts its contents would not be of interest, the label alone being sufficient for many purposes.

4.3.3 Definition of a system and the first principle of systems

With the above descriptions of systems engineering and systems in place, an improved definition of a system can be provided. A system can be defined only in the context of its environment (Hitchins, 2003). The environment is defined as the system in which this system is contained as well as all other sibling systems with which this system interacts. Based on this view, the following definition of a system is formulated:

A system is an open set of complementary, interacting parts with properties, capabilities, and behaviors emerging both from the parts and from their interactions.

This definition of a system leads to the first principle of systems:

The properties, capabilities, and behaviors of a system derive from its parts, from interactions between those parts, and from interactions with other systems.

Furthermore, Hitchins (2003) provides a corollary to the first principle:

Altering the properties, capabilities or behavior of any of the parts, or any of their interactions, affects other parts, the whole system, and interacting systems.

Hitchins states that this corollary may be obvious, but the implications are widely overlooked. Failure to understand systems or deliberate choices to overlook the implications has throughout history been the reason for many accidents and catastrophic events.

Some fundamental system constructs can be identified (Hitchins, 2003) that are fundamental to all kinds of systems and that can be useful in understanding and applying a systems engineering oriented approach to the design of systems. All elements of a system are interconnected directly or indirectly and in interaction. Connectivity and relationships are fundamentals that enable interactions. Conservation laws apply, except for information flows. The parts and their interrelationships of any system form patterns referred to as the configuration. The configuration can be more or less disordered (see section 4.3.7 on complexity). The architecture, related to configuration, is fundamentally the structure created by grouping and linking parts to form interacting sub-systems and systems. Containment is the demarcation of a number of complementary parts within a set boundary, to form a system or sub-system. Complementation is the ability of parts of a system to complement each other to "fit" with the others such that all mutually contribute through interactions to the whole system. A full set of complementary parts make a complete whole. Hierarchy is the vertical structure formed when complementary parts form a complete whole that can be considered as a unit with its own properties, capabilities, and behaviors. If that unit complements other units then a higher level of hierarchy emerges. Hierarchy is indicated by emergence, and vice versa. Emergence is the phenomenon of properties, capabilities, and behaviors evident in the whole system that are not exclusively ascribable to any of its parts. A classic example is self awareness in the human brain. Emergent properties generally refer to attributes observable from the "outside" of the system. Emergent capabilities generally refer to limits of functional ability. Emergent behaviors generally refer to responses to stimuli. In every case of emergence, the source is interaction between the parts sometimes, as with the brain, very many parts so that the complexity of the phenomenon defies simple explanation.

4.3.4 Holistic, synthetic, and organismic

The three terms holistic, synthetic, and organismic are according to Hitchins (2003) fundamentally important characteristics of the systems created using a systems engineering approach. Hitchins states that the systems created are holistic in concept, synthetic in design, and organismic in structure. Holistic means that the systems should be conceived, designed, and developed as a whole and with consideration to their interactions with other systems in their environment. Synthetic refers to that the system is built of parts that are themselves systems and that these are interconnected in such way that the system as a whole will deliver the appropriate emergent properties, capabilities, and behaviors. Organismic is that the system is viewed as an open system and analogous to an organism. The various elements of the system are interactive and mutually interdependent in such a way that a constraint on the whole system necessitates complementation and compromise by the elements themselves and their way of interacting.

4.3.5 Elaboration and encapsulation

An interesting differentiation between elaboration and decomposition is given by Hitchins (2003) who points out that the major difference between the two is that, in an elaboration, all elements and their interactions remain. Typically, the interactions among the elements are lost in a decomposition, which is regarded to be a reductionist approach in this sense. The purpose of elaboration is to provide more details about a system like a magnifying glass that provides more details about the elements without taking them apart. Decomposition is more like a divide-and-conquer approach where elements are broken down to be dealt with individually. Both context and interaction may get lost in such an approach. Hitchins writes: "Synthesis is the opposite of reduction. Reduction looks into a system; synthesis looks out of a system. Reduction breaks down; synthesis builds up. Analysis, looking into things, yields knowledge; synthesis, looking outwards, gives understanding."

While elaboration can be described using the analogy of a magnifying glass, encapsulation can be viewed as the placement of containers around sets of entities. The container provided through encapsulation can then be treated as a single entity that can represent the contained set of elements including their emerging properties, capabilities, and behavior. The containment using encapsulation does not affect the interactions that remain intact. The dual processes of elaboration and encapsulation are illustrated in Figure 26.

Figure 26: Elaboration and encapsulation (Hitchins, 2003)

4.3.6 Systems engineering philosophy

Systems engineering focuses on more generalized, as opposed to specialized, design solutions in order to avoid making premature assumptions that may lead to sub-optimizations and/or missed opportunities. Design solutions are therefore formulated with a high level of abstraction which opens a generation of many options. Identification of many options as well as many criteria for their existence leads to a questioning of many otherwise tacit assumptions. Hitchins (2003) refers to this as disciplined anarchy. Complementary to this wish to maintain a high level of abstraction is the advocacy of breadth before depth. The whole problem space should be explored before examination of the details of parts of the potential solution. A consequence is that there is a preference in systems engineering to explore one level at a time in order to avoid the risk that some parts of the system receive too much focus and attention or face the risk of being neglected. Elaboration of purposeful systems necessarily puts a focus on functions and activities that can be organized and grouped into various sub-systems. Consequently, systems engineering advocates a functional before physical approach. A further characteristic of systems engineering is that it is throughout the process concerned with emergent properties, capabilities, and behavior. This outward looking approach that characterizes synthesis is fundamentally important for managing complexity, creating non-linear systems, and achieving system goals. Systems engineering, then, is not engineering in the conventional sense that relates only to machines and other manufactured things. It is, however, engineering in the sense of the planning and bringing about of something with ingenuity (Hitchins, 2003).

According to Ackoff (1981) to resolve a problem is to find an outcome that is good enough, one that satisfices; to solve a problem is to select an approach that yields the best of all possible outcomes, one that optimizes; to dissolve a problem is to change the nature of the problem, or its environment, such that the problem disappears. Ackoff proposes that to resolve is a qualitative rather than quantitative approach, relying on past experience and current trial, and is rooted in common sense and subjective judgment. To solve a problem, on the other hand, is a research approach based on scientific methods, techniques, and tools, such as mathematical models and simulations. To dissolve a problem is to idealize, rather than to satisfy or to optimize, and to change the situation so that the problem cannot or does not arise. In practice, dissolving would mean not addressing the scenario as given: for instance, one way of dissolving the problem would be to change the environment (i.e. a new scenario).

4.3.7 Complexity

Hitchins (2003) discusses a theory of complexity and states that complexity is subjective. Familiarity with something tends to reduce its complexity (or the perception of complexity). Hitchins (2003) views complexity as being generated by three factors in combination: variety, connectedness, and disorder. Variety in this context is seen as a commodity that one can add or take away without having to specify what constitutes the variety. Adding variety is adding a new category. Connectedness, or connectivity, is the degree of linkage between elements. The third factor contributing to the perception of complexity is disorder. Disorder can be thought of as tangling. If the various elements of a system are arranged in such way that various linkages cross, interweave, and overlap the resulting tangle of interconnections increases the perception of complexity.

Figure 27: Perception of complexity (from Hithcins, 2003).

The perception of complexity is illustrated in the two leftmost views in Figure 27. The degree of variety is unchanged in views (a) and (b) (i.e. they are composed of the same set of elements and linkages). However, the perceived complexity is very different. Since the variety is unchanged, this perception must be due to the untangling of the interconnections that was achieved by rearranging, or reconfiguring, the elements of the system. From this follows that the perception of systems within a complex set of many interconnected parts is encouraged and simplified by reconfiguring the elements so as to reduce the overall entropy. An approach to do this would: 1) evaluate the configuration entropy; and 2) reconfigure the pattern to minimize the disorder. The right part of Figure 27 uses an N2 chart to represent the system and to calculate the configuration entropy. Hitchins (2003) uses N2 chart representation to calculate the configuration entropy. Let ND be the number of linkages with distance D where D is the distance of the linkage from the leading diagonal in the N2 chart. Then the configuration entropy is the sum of all multiplications of ND x D in the N2 chart (for additional details see Hitchins, 2003). The configuration entropy of the system in Figure 27 is calculated to be 104 with maximum entropy of 572, which yields an entropy ratio of 0.182 for this configuration of the system.



Continue ...

Go to table of contents.