Zef Damen Systems Theory

Introduction
Basics
Functional decomposition
System dynamics
Other characteristics
Systems theory and logistics


Introduction

Both, Flexible Mail Processing and Service Controlled Agile Logistics (SCAL) are based on Systems Theory. In stead of only analysing the world of logistics, to know what can be found "out there", it tries to understand why all these things are there, and what things should (or should not) be there. In stead of observational, it follows a more deductive approach: a reasoning from the general to the particular, in which conclusions follow more or less necessarily from given premises.

(Ideas and definitions are mainly taken from: In 't Veld, Analyse van organisatieproblemen, een toepassing van denken in systemen en processen (Analysis of organisational problems, an application of thinking in systems and processes); Elsevier Amsterdam/Brussel, 1983, pp. 8-27 (In Dutch)).


TOP

Basics


Elements

A system is a set of elements distinguished by the systems researcher within all of reality according to a certain goal stated by the researcher. These elements are related to each other and to other elements of total reality. Elements are the smallest parts of the system the researcher wishes to recognize. They can be material or non-material. All elements have certain attributes. When an attribute is quantifyable, it has a value. Sometimes, attributes are only qualifyable.

Relationships

Between elements there exist relationships. These denote certain coherence or interaction between the elements. That means that values of attributes of one element depend on values of attributes of other elements. And vice versa. The relationships with the outside elements determine the purpose or goal of the system. Since these relationships depend on the system's boundary, this boundary is chosen not at random, but purposefully. The set of related outside elements is called the system's environment. These relationships describe the interaction of the system with its environment: the way the system is influenced by its environment and influences its environment. If there are no influences, no relationships with the environment, the system is said to be closed, otherwise, it is called an open system. For our deductive approach, we need an open system.

Function

Therefore, our system has a purpose, a goal. It is the system's function to realize its purpose, to reach its goal. Purpose or goal are relationships – with the environment –, the function of a system (or, as will be seen in a moment, of a subsystem) is a responsibility. Here, we will treat the term function in this strict way.

TOP

Functional decomposition


Subsystems

An important characteristic of a system is the possibility to subdivide it into subsystems. A subsystem is a subset of the elements of a system, such that all relationships between these elements and with the other elements within the system and with elements outside the system remain unaltered. It therefore consists of interrelated elements and has relationships with elements outside the subsystem (within or outside the original system). A subsystem has all the characteristics of a system; thus, a subsystem is a system. A subsystem is smaller than the original system (it has less elements and therefore fewer relationships). It is always possible to iteratively subdivide an arbitrary system into subsystems, until the elements are reached.

Decomposition

Systems theory is used to learn to know an arbitrary system in all its relevant details. This may be done to understand how a certain system is working. Or it may be necessary to be able to realize a system in practice. It is accomplished by starting at the system as a whole. This is called the trivial level. The purpose of the system is defined as precisely as possible, in terms of the relationship with the environment (the environment has to be defined). This is equivalent to defining the system's function. Next, iterative steps are undertaken to subdivide the system into subsystems. This is not a trivial task. (In a system consisting of N elements, 2N - 1 subsystems exist). To guide this subdividing process, a hierarchy of subsystems is created. At each next level of the hierarchy, the total system is given, every time in more detail. At each step, the system is broken down into different parts along boundaries of weakest coherence, the boundaries of the subsystems with least dependence on other subsystems. These are the "natural" boundaries. Since the relationships "crossing" the boundaries determine the function of the subsystem, looking for the weakest boundaries is equivalent to searching for subfunctions that are as independent as possible. Therefore, building the hierarchy of subsystems is guided by looking for "main" functions (within broader functions), those that are maximally independent. This is called functional decomposition.

Top down versus Bottom up

When functional decomposition is started, it might be unclear, what exactly the elements are one is looking for. Dependent on the reason for the system's analysis, the level of the elements may differ and therefore the "stop criterion". If the analysis serves a closer understanding of an unknown system, functional decomposition is stopped when a sufficiently detailed level is reached. If on the other hand the "top down" approach of system's analysis is used to be able to realize the system in practice, functional decomposition should reach a level at which the subsystems are known to be realizable in one way or another. This requires an understanding of how parts of the system could be built, thus knowledge of the "bottom up". In both cases, it may be hard to tell beforehand where (when) to stop. Therefore, after each step of building the mentioned hierarchy, one has to judge whether enough detail has been gathered and if the iterative approach will be continued or stopped.

TOP

System dynamics


State

In general, systems will not be invariable, neither in the values of the elements' attributes, nor in the number of elements or in the mutual relationships or in the relationships with the environment: systems will change over time. The way systems change is very important. It is how the environment will conceive the system, and how the system reacts upon its environment. When we think of input as the "influences" from the environment onto the system and output as "influences" from the system onto the environment (we will define input and output more precisely in a moment), then we can define the state of the system as the set of elements of the system and the values of their attributes, together with the input of the system at a certain moment that uniquely determine the output of the system at the next time interval. This state changes from moment to moment, partly influenced from the outside by the environment, partly due to activities of elements within the system itself. The total range of elements and values that the state of the system can reach is called the state space.

Dynamic systems

Systems that change state are called dynamic systems. Most of the changes of the system's state are caused by actions of the system's elements. Not all elements, however, have to be able to cause the system to change state; those that can are called active elements, those that cannot are called passive elements. The change of state is called an event. Whenever (as long as) an event triggers another event, this is called an activity. Therefore, an activity is a series of changes of the system's state until the system reaches a steady state, as long as the input of the system does not change. The next change of state is caused by changes in the input of the system.

Process

In general, active elements are part of the system and passive elements flow through the system, from input to output. The system will transform these elements while they flow through. From this viewpoint, a system can be regarded as having an input, a throughput and an output. The input consists of elements flowing into the system (these might be active elements) and the relationships from outside that determine the values of attributes of system's elements; and the output consists of elements flowing out of the system (these also might be active elements) and the relationships from the system's elements that determine the values of attributes of elements of the system's environment. A process is a series of transformations during throughput, by which the input elements are changed (the values of their attributes are changed), and transformed into output elements. We call the way the system's output varies over time its behaviour. When this behaviour relates directly to the system's input and not to any previous situation (state), it is called static behaviour. If, however, it depends not only on the input, but also on the state of the system, it is called dynamic behaviour. In this case, the system's state must remember relevant history.

TOP

Other characteristics


Structure

All the relationships together form the structure of the system. The word structure is often used, also in different contexts; it is always important to be clear about what kind of structure is intended. Also, the structure of a system can change.

Purpose, function, task

There should be made a clear distinction between purpose, function and task(s) of a system (or subsystem). We now can state the purpose of a system as the output (or better behaviour) it should have. It is its function to realize it. To fulfil this function, it has to perform its task: have its elements carried out activities such that the wanted output is generated. So, the purpose as well as the function are given in terms of the output (elements and relationships), the task is given in terms of activities, series of events that change the system from state to state, carried out by the system's active elements. Purpose and function have to do with results, the task refers to the way to achieve these results. The main confusion between purpose, function and task is created by functions gradually changing into actions when detailed step by step, which is exactly the purpose of functional decomposition. Notice the difference made here between actions and activities.

Aspectsystem

Besides being able to divide a system into subsystems, it is also possible to divide it into aspectsystems. An aspectsystem is a subset of those elements of a system related to each other according to a subset of all the relationships, such that all elements and their relationships within or outside the system remain unaltered.

TOP

Systems theory and logistics


Transport Unit

With logistics, two things are most fundamental: goods and their displacement. Without goods, there does of course not exist any goods distribution. The kernel of the problem is that these goods are present at a certain place at one moment, and someone (often the owner of the goods) wants them at another place at a next moment. Due to physical reality, the goods have to be moved. Since this movement often affects long distances, appropriate vehicles are necessary. In Flexible Mail Processing, the Transport Unit (TU) has been introduced as an abstraction of both, goods and vehicles. It serves as a "placeholder" for anything that has to be moved (transported). The TU is a passive element. When it stands for a package a customer wants to be delivered at some address, this TU originates from outside the system and flows into it (input). Other TUs are created inside the system itself: pallets, trucks, trays, roll containers, etc. The package delivered at the requested address still is a TU, this time flowing out of the system (output).

Basic network

An important characteristic of the TU is that it can contain other TUs: the TU is a hierarchical entity. Compare a truck loaded with pallets with packages: three levels of TUs. These higher level TUs are mainly created and dismantled within the system. We therefore need some subsystems that transform TUs: build new, higher level TUs from lower ones, and dismantle higher level TUs releasing the lower ones again. And, of course, we need transport subsystems. Transformation is carried out at geographically fixed places, transportation connects these places. TUs, transformation and transportation together form the basic network. The transformation places are the nodes, the transportation connections the arcs. TUs flow through the network and from and to the environment.

Shipments

Sometimes, customers are interested in sending goods that belong together. This is called a shipment, and cannot be denoted by a single TU. Therefore, a further distinction must be made between goods (input or output) and TUs. Goods are "translated" into shipments consisting of one or more TUs as soon as they enter the system. Shipments may be viewed as logical entities, whereas TUs are references to single physical units. As a consequence, a TU must be able to reference both, other TUs belonging to the same hierarchy and other TUs belonging to the same shipment.

Control

Both, transformation and transportation rely on active elements; they carry out activities. For the network being able to perform distribution services, it has to accept TUs, transform these into appropriate "intermediate" TUs, at least transport these, and deliver the original TUs. Transforming and transporting TUs should be done in a coordinated way so as to guarantee quality requirements agreed upon. The active elements need to be controlled in one way or another. This is the function of a control subsystem. It should have an "idea" of the service requirements of the distribution services to be delivered and will have to translate this into the right "activation signals" to the active elements of the basic network. Although control may be viewed as a single function (at a high level of abstraction), it is not intended to be centralized. On the contrary, it most probably is scattered over the entire network, even over independently operating parties. (Notice, that being centralized or distributed has nothing to do with the control function; it is a way of realizing it in practice). So, the control subsystem receives service requirements constraints from "above" and will have to translate these into control information and send this information "down" to the basic network. At its own level, it has to manage all the coordination among the active elements and their actions. Service Controlled Agile Logistics (SCAL) is a control subsystem in which the customer is at the top of this control chain, and really has full control.

TOP


Copyright © 2000-2002, Zef Damen, The Netherlands.

Go back Last updated: 2-May-2002


Since 1-February-2005