System Concepts Pengantar Teknik & Sistem Industri Jurusan Teknik Industri ITS
Definition A system is an organized assembly of components. Organized means that there exist special relationships between the components. The system does something, i.e. it exhibits behaviours that are unique to the system. Each component contributes towards the behaviour of the system and its own behaviour is affected by being in the system. No component has an independent effect on the system. Groups of components within the system may by themselves have properties. The system has an outside an environment which provides inputs into the system and receives outputs from the system. The system has been identified by someone to be of special interest for a given purpose.
System Boundary The separation between the system and its environment means that there is a boundary. In fact, boundary selection is the most critical aspect of systems thinking. Out there & inside us view
Systems as Black Boxes The complexity of real life may be such that t we have no or only incomplete knowledge of the inner workings of a system, even if we are able to identify the physical components. Often the major reason for this lack of knowledge is that thesystem behaviour is affected by random aspects. In other cases, the relationships between components are only partially understood In other situations the transformation process is known exactly. However, rather than represent it in full detail, it may be adequate to view the inner working as a black box and simply express the various activities of the transformation process by a single functional relationship
Hierarchy of Systems
Different Kinds of System Discrete systems (a discrete system changes its state at discrete points in time) The number of telephone lines or the number of operators busy In a predator/prey system, the state is described by the number of predators and number of prey alive at any point in time In the loom repair system, two state variables of prime interest are the number of machines operating and the number of machines broken down
Continuous System (the state of the system changes continuously) Many industrial processes, particularly in chemical and petrochemical plants, should be viewed as continuous systems. The process used by warm blooded animals to maintain the body temperature within a narrow range is also a continuous system.
Deterministic & Stochastic System If the behaviour of a system is predictable in every detail the system is deterministic. For example, the solar system, animated neon advertising signs that go through a regular pattern, a sequence of traffic lights along a one way street If the behaviour of a system is not completely predictable and affected by random or stochastic inputs, the systems are called stochastic systems.
Closed and Open Systems A closed system has no interactions with any environment. No inputs, no output. In fact, it has no environment. In contrast, open systems interact with the environment, by receiving inputs from it and providing outputs to it. In real life there exist no truly closed systems.
The Steady State of a Probabilistic System Stochasticsystems in the long run tend to approach a state of equilibrium, also called a steady state. This state of equilibrium is independent of the state the system starts out from. Few stochastic systems ever reach their state of equilibrium. Small or large random disturbances, e.g. a sudden random surge of emergency calls, or a severe storm in an ecological system, may disrupt system behaviour and push it away from these long run averages. But tin each case, the system will gradually approach the same or a new state of equilibrium again.
Feedback Loops Feedback kis a common feature of most systems, both human activity systems and natural systems. They often are the main cause of complexity. Feedback can act positively or negatively. Positive feedback increases the discrepancy between the future state of the system and some reference state, such as an equilibrium state or a desired target state. tt In other words, the system state tt tends to deviate more and more from its reference state. In contrast, negative feedback decreases the discrepancy between the future state and the reference state
Control of Systems Three conditions are needed ddto exercise control over system behaviour: A target, objective, or goal for the system to reach. For a deterministic system this may be a particular state of the system. For stochastic systems it may be a desirable steady state. A system capable of reaching the target or goal. The difficulty is that for stochastic systems there may be no way of guaranteeing thatt this goal is ever reached. hd Some means of influencing system behaviour. These are the control inputs (decisions, decision rules, or initial states). How these control inputs affect system behaviour is an important aspect of studying systems.
Three Types of Control Open Loop Control Closed Loop Control Feed Forward Control
Open Loop Control Inputs imposed on the system based only on the prediction of how the system behaviour responds to them. No account is taken of how thesystem actually responds to the control inputs.
Closed Loop Control Information about the system behaviour, possibly in response to previous control inputs, is fed back to the controller for evaluation. This may lead thecontroller to adjust the control signals. There are two mechanism: Feedback control Self regulation
Feed Forward Forward Control Reacts to changes in some critical state variables or outputs. Rather than react to events after they have happened, a feed forward control mechanism predicts how changes in inputs (uncontrollable or controllable) are likely to affect system behaviourand and thensendscontrol signals that will maintain system behaviour as closely as possibleon on the desiredcourse course, thereby counteracting the effects of input disturbances
The Problem Situation Pengantar Teknik & Sistem Industri Jurusan Teknik Industri ITS
The Problem Situation The context within which the problem occurs. It is the sum or p aggregate of all aspects that can or may affect or shape the problem or issue of concern.
Six Elements of a Problem The decision maker The decision maker s objectives The associated decision criterion The performance measure The control inputs or alternative courses of action The context in which the problem occurs
Objective & Decision Criterion An objective is the end towards which h effort is directed,an d aim, goal or end of action an ambulance service wants to find the best location in a small city, where best is interpreted as reaching any emergency as quickly as possible Criterion isdefined as theprinciple orstandardonwhich on a judgment or decision is based. Both principle and standard imply a rule. So a criterion is the rule used for judging whether or how well the objective has been achieved to minimize the sum of all times to reach every road location in town, to minimize the sum of the squared times, to minimize i i the maximum time bt between the ambulance service and any locations in the city.
Exercise 1 Define the six problem elements for Emergency call centre problem Vehicle routing problem
Stakeholders The decision maker, other parties affected but without any control over the situation, and the analyst Categories: The problem owners The problem users The problem customers The problem analysts or solvers
Exercise 2 Define the stakeholders for Emergency call centre problem Vehicle routing problem
The Aids to Depict a System Goals: Acquiring a sufficiently complete and detailed understanding di of the problem situation ti for a successful system intervention Gettinga thorough feel for anything that may impact on the outcome Diagrammatic aids to depicta system: Mind maps Rich picturediagrams Cognitive maps
Mind Map When you think about something a phenomenon, an issue, or a problem a host of thoughts are evoked in your mind: things, aspects, andconcepts concepts, including fears and aims, data and facts, and the possible actions and reactions by yourself or other people p or entities involved and their consequences, both planned and unplanned, desirable and undesirable, that result from such actions, and the wider context or environment of it all. A mind map is all this (or a judiciously chosen subset) put down on paper in headings, slogans, or sentences
The things are arranged in a meaningful way by showing aspects closely related in groups, by lines that connect things which are related, and by arrows that indicate causal relationships between items. No formal conventions are used.
Rich Picture Diagram Rather than show the various aspects in words or short sentences, P. Checkland [1993/99] suggests drawing a cartoon like pictorial summary of everything (or almost everything!) the observer knows about the situation studied. Note that term rich picture does not, in the first place, mean a drawing. It is simply a more colourful term for a situation summary. Its cartoon like representation is called a rich picture diagram. However, this is rather clumsy and long. So if it is clear from the context, we will refer to the diagram simply as a rich picture.
Cognitive Diagram Cognitive mapping is a tool lthat t C.L. CL Eden [1983] adapted dfrom G.A. GA Kelly s (1955) personal construct theory. In contrast to mind maps and rich pictures, which are suitable to represent an individual s personal as well as a group s aggregate perception of the problem situation, a cognitive map only captures the subjective, personal perception of an individual. It takes the form of a network of statements, expressing concepts ideas, goals, concerns, preferences, actions and their contrasts or opposites. The concepts are linked together by arrows, which indicate the direction of connections, i.e. which concept leads logically to which other concept(s). p() Cognitive maps have some similarity to mind maps that capture means end or cause and effect relationships. (They are also related to causal loop diagrams)
System Model A system model lis a representation ti of all essential parts of a system A model may be: Iconic Analogousor or symbolic Mathematical Characteristics of a good model: Simple Complete Easy to manipulate & communicate with Adaptive
Some Examples Causal Loop Diagram
Influence Diagram
Material/Information Flowchart
Precedence Chart
Fishbone/Spray Diagram
Fault Tree Diagram
Decision Flowchart
Systematic Model CONSTRAINTS DECISION INPUTS PROCESSES OUTPUTS FEEDBACKS