Bruger:Erikskeel/Sandkasse

Introduction Reality and models consists of a never ending spiral of innovative technologies that can be modelled up to a speci�c level of uncertainty. Companies often take a longer time to model a manufacturing line as to get a CAD model running. Throughout the course, the main objective was to understand modern technologies that support innovation along with product and service development. Before understanding technologies available, one must understand the di�erence and similarity between theories, models and reality. All three concepts interlink with each other but have very contrasting connotations. 1.1 Theory Theory is a big topic that is best described using the work of Whetten [1] (1989). Theory is determined by asking four aspects `what', `how', `why' and a fourth group of `who', `where' and `when'. These four groups are interlinked and an example can be shown of each: What are the variables or concepts used in theories? This basic question is essential because it asks the variables what they have to do with the theories. The theories have to relate to a reliable source or they will not be recognised. How do these variables or concepts relate? When multiple concepts are used, they have to collaborate to make a point to the reader. Theories are individually unstable and need two or three to determine a plan. Why are these theories relevant? Theories need to be relevant to the text or why would they be placed into a project. It also explains the nature of the relationship between the variable and the concept. This means that the what and why are interchangeable if written correctly. What and how are descriptive, while why is the explanation of the relationship between what and how. To summarise the �rst three elements that are used here, one cannot simply use one at a time because they all describe di�erent aspects of theory. Theory has the power to predict models and simulate reality but contains many limitations and con- straints. The fourth group then justi�es the circumstances that the theory is applied by asking; who does the theory apply to; where does this theory apply; when does the theory apply. These questions are im- perative because they ask speci�cally what �eld the theory belongs in. For instance, it could apply only for quantum physics but does not aid in business economics. Moreover, the relationship between variables in a theory might be time constrained because the theory may only apply for a logistic time schedule. [2] To summarise Whetten, it must not only include the what and how elements but also logical reasoning to explain the relationship. Logical explains the good reasons behind the why and reasoning because your logic must be based on prior learning or experiences. Moreover, theory is important because it is widely misused word that needs to be explained before implicitly using in our everyday vocabulary. This leads us onto a model which must be based on a theory. Box (1976) noted that \all models are wrong, but some are useful." [3] 1.2 Model Models are used to virtualise reality. It does not have to capture all attributes of a model but does have to represent the most relevant aspects. A model is created to aid further improvements on prototypes but also to show potential clients and stakeholders a design before o�cially bought. [4] Model-Based-System- Engineering (MBSE) is an academic form of modelling a system required and its progress through the later life-cycle of the product. There are multiple advantages of creating models instead of reality; �rst, the cost is reduced as most �rst prototypes need improvements and the size can be reduced or increased depending 3 on the model (creating a house is di�cult but on a model, it can be designed e�ciently and e�ectively on CAD). There are two types of models: theoretical modelling and physical models. Before an engineer can build a full-scale model, one must make a physical model �rst and from that decide how the model can be adapted. One of the other hand, a scientist will take the theoretical model approach �rst as an idea must be simpli�ed and/or idealised before making a physical model. [5] Models are make-believe items that Walton explains as a children's game. [6] He says that if tree stumps are props and the children agree that they count as bears. This principle and prop combined make a �ctional proposition. To sum up, the stumps in the children's game are not representations but a work of a �ctional model. The reason this is a good example is that a model does not have to represent the �nal product, but as long as a theory is placed along with it, then it becomes closer to reality. 1.3 Reality Reality is the the known and the unknown, described by models, formulas and theories. Reality is described by a common agreement, and no single person can change or de�ne reality. Reality uses theory and models to manufacture an architectural physical model that can be manipulated by hand. Reality is in uential because it brings together both the terms theory and model into one design. [8] This one design then becomes what we call reality, so it is physical and in theory, works as the model and simulation both did as well. Reality in the sense of our project meant that the robot could function through the course, with the same precision intended. Reality, although seen by the naked eye, can be perceived di�erently as shown in the Reality and Models lesson with illusions. This is important to factor into projects as videos, although they perceive reality, can be altered to shift the sense of accomplishment to a person. In this report, we will attempt to show through pictures and virtualisation that our reality matches that which can be seen on simulations. Therefore reality will not be delusional. To do so, simulations must be understood with the mathematical kinematic formulations before forces can be supplemented. Reality is hard to de�ne because we only know reality to a certain extent. Increasing sample sizes and ensuring relevant information is chosen allows us to get close to reality but never quite catch it. [8] 2 Principles of Modelling and Control Simulation is de�ned as `the imitation of the operation of a real-world process or system over time. The act of simulating something �rst requires that a model be developed; this model represents the key characteristics or behaviours/functions of the selected physical or abstract system or process. The model represents the system itself, whereas the simulation represents the operation of the system over time.' [9] In simpler words, a simulation can be represented by a model as long as the model is accurate. One has to remember that the real world is di�erent to a simulation, no matter how accurate, which will be discussed later in this section. There are three areas of Mechanical Modelling which will be examined: Kinematics, Statics and Dynamics. Kinematics is the study of a motion in a mechanical structure (does not include any forces). Statics is the study of the equilibrium upon a system because of forces acting against each other. Dynamics is the study of motion as a result of unbalanced forces that act on a kinematic structure. In order to describe each area on a mechanical model, it has to be exempli�ed. 2.1 Kinematics Before studying the dynamic behaviour or statics, a model must �rst be created to show the kinematics of the system. The most elementary is so called Kinematics of the point where a point can be shown via 4 vectors or coordinates of the vector. These representations of the position of a point can be converted into three formulations: Cartesian, Complex plane formulation and Polar Formulation. An example of kinematics of a point is: Figure 1: Vector P on a plane Image 1 can be described �rstly by using vector ^P= (^P - ^ O). If this is converted to a Cartesian equation then the x-axis is replaced by i x-axis and the y-axis replaced by j y-axis. Therefore ^P(t) = ix(t) + jy(t). The complex plane can then be formulated. A complex plane contains both an imaginary and real part, hence the name complex. The imaginary is the y-axis and real is x-axis, so it can be shown on the same graph as above like this: Figure 2: Complex plane with vector P Im stands for imaginary and Re is Real. The same point on this plane is used but written in the com- plex plane formulation: ^P = Px + i Py = ^Pcos � + i ^Psin � . Once the complex plane formulation is found, it is easier to place into the polar formulation Pe i � . It is generally known that it is possible to pick and choose any of the position point formulations as they represent the same. This is all great if the robot was static and the position was needed to make a model but this rarely happens. The project's requirements include a moving robot that can carry objects so the velocity at a given point can be calculated. In the real world, it is performed on a rigid body. A rigid body is de�ned as a body which the distance between two points never changes, so it does not deform under the in uence of forces. It can be very handy to be able to analyse the movement of the robot and all the points within the rigid body. This is because we can visualise and process the expected movement within every point of the rigid object. With that, we are 5 able to calculate values like velocity and acceleration. This helps us to understand how external forces are going to a�ect the rigid body but also when the movements of several rigid bodies rely on each other. When talking movements of the rigid bodies we look at three di�erent ways of moving. First we have the translational movement. This is very similar to pushing a box. The rotation is �xed and therefore the velocity vector is the same in in absolute value and direction. Next we have rotational motion. This is when the rigid body has one point (can be both inside and outside of the rigid body) of zero velocity, which is called the instantaneous centre of rotation. The last one is called roto-translational and is a combination of the two mentioned before. Another important thing for the understanding of the movement is how rigid bodies can be connected. There are di�erent kinds of joints that have di�erent kinds of limitation with respect to degrees of freedom. These are called kinematic constraints. Once we understand how the rigid body is moving we can use this knowledge to �nd the acceleration, velocity and position that we want for a speci�c point. This can be very useful in the real world as it bene�ts our prototype developing. When visualised one is able to eliminate mistakes and make changes, in our case with the robot. This means that we are able to optimise the design and production of the robot itself. Figure 3: Shows the di�erent kinds of joints where the red arrows is movement allowed and the red arrows describe a reaction for or torque. 2.2 DoF Degrees of Freedom (DoF) refers to the number of independent coordinates that have to be speci�ed in order to determine the speci�c location of a body. If the body is a point mass then only three coordinates are needed to determine its position (x,y,z), but if the body is extended such as a spacecraft then three angular coordinates are also required (x', y', z'). 2.3 Dynamics and Statics In this next section, dynamics and statics will be explained which can explain how a physical system can be represented by mathematical equations. The basic notions before simulating dynamic objects, are Newton's Three Laws of Motion. The �rst being the Law of Inertia: If an object is at rest, it will stay at 6 rest unless acted upon by an unbalanced force. The same can be said about an object in motion, unless it is acted upon by an unbalanced force, it will continue in the same direction. Newton's 2nd Law states that the acceleration of an object is proportional to the magnitude of the net force, which is indirectly proportional to the mass of an object. It is written as F = m*a where F is force, m is the inertial mass of the object and a is acceleration. The 3rd and �nal law is: For every action, there is an equal and opposite reaction. Forces generally depend on three factors: The position of the point (r), the velocity of point P (v) and time (t). Friction is the third and �nal part of simulation, which is known as a dissipative force. This is when two surfaces are in contact, one being dynamic and the other being static. Friction is caused normally by the chemical adhesion between the elements of the two surfaces so it can dramatically change depending on the materials used. 2.4 Control In order to achieve the behaviour we have simulated, there is a need to minimise the di�erence between the the theoretical model expected output and the output that is actually measured in reality, the cause for the di�erences in output has to do with factors that were not taken into consideration at the time accounted for in the chosen model, those factors are explained in the following section 2.5 (Challenges faced creating simulations). In order to compensate for the di�erences in output, we use control loops. There are 3 types of control methods that can be applied to a process: 1. Feed-Forward Control 2. Feedback Control 3. Proportional Control Feed-forward control is a purely model-based approach which describes an element within a control system. It passes a controlling signal from a an external source to control the input of a process in order to to improve its output by compensating for a known disturbance. In feed-forward control loops, the signal that passes is usually a product of mathematical models, based on knowledge about the process, and would be used when the Steady state error (The constant error between the target value and steady state output) is known. This requires perfect knowledge of the model parameters and the applied load. Another consideration to whether or not to use feed-forward control depends on whether the degree of improvement in the response to the measured disturbance justi�es the added costs of implementation and maintenance. The economic bene�ts of feedforward control can come from lower operating costs or increased quality consistency. [11] Being purely model based, there are limitations to the use of feed-forward control such as: • E�ects of disturbance or command input must be predictable • Will not be accurate if the system changes In order to counteract disturbances that come from sources that are not included in a model, which can be seen as the di�erence between model and reality, feedback loops can be used. 7 Figure 4: Block diagram of feed-forward control loop Feedback control is a sensor based approach describes a control system that counteracts a disturbance by manipulating the input variable according to changes in measured output. Feedback control process can be described as Sensing - Computation - Actuation. In order for disturbance correction to take place, the a�ected output variable must �rst a�ect the controlled input variable, meaning that the correction is made in retrospective hence the term `feedback'. Figure 5: Block diagram of feedback control loop In feedback control, we compute the error function which the controller tries to minimise as the dif- ference between the reference point (or `set point') which is our desired output and the measured output value and, and then manipulate the control input according to a function of the computed error. By using feedback which is a sensor based allows the control system to: • Reactive / Error-driven • Automatically compensates for disturbances (controller acts on error) • Automatically follows change in desired state (set point can change) Showing that feedback is adaptive compared to the model based feed-forward control system, as it can account for instantaneous changes to the process, while a feed-forward control system depends on the variables being constant. 8 2.5 Proportional control When the control response applied to the system is proportional to the error, we describe the system as a proportional feedback controller (P Control), the proportional control address the current system behaviour. In cases where a proportional feedback controller is applied, oscillations may occur due to ex- ternal loads; or harmonic oscillations that might be generated in a mechanical system by the proportional feedback controller when it attempts to compensate for the error if no dampening factor exists. Note that the proportional response can be adjusted by multiplying the error by a constant Kp . In order to reduce the oscillations of the system, a derivative factor (given as Dout = Kd de(t) dt ) can be added which takes the derivative of the error and acting as a dampening element on the mechanical system. Determining the slope of the error over time and multiplying this rate of change by the derivative gain Kd predicts future system behaviour and thus improves settling time and stability of the system, this is known as a proportional-derivative control (PD Control). In order to achieve even better performance, the past system behaviour can be taken into account, the integral factor is the sum of the instantaneous error over time and gives the accumulated o�set that should have been corrected previously (Iout = Ki R t 0 e(T)dt ). Introducing an integral factor is used to add long term precision to a control loop. It is almost always used in conjunction with proportional control (PI Control), as controlling a mechanical system using integral control alone will oscillate with bigger and bigger swings until something hits a limit. Combining the three factors above leads to the �nal controller, the PID controller which is commonly used in industrial control systems : \PID stands for proportional, integral, derivative. These three terms describe the basic elements of a PID controller. Each of these elements performs a di�erent task and has a di�erent e�ect on the functioning of a system. In a typical PID controller these elements are driven by a combination of the system command and the feedback signal from the object that is being controlled (usually referred to as the \plant"). Their outputs are added together to form the system output." The PID controller's response is de�ned as Pout + Iout + Dout = PID, equals the combined term: Kpe(t) + Ki R t 0 e(T)dt + Kd de(t) dt and can be described by the following block diagram: Figure 6: Block diagram describing a classic PID controller 9 By implementing a PID controller, [10] there becomes a close correlation between the error and the controller output which results in quality system output with minimised error over the changing time and load. Note that achieving the correct correlation requires tuning the controller by adjusting the gains Ki, Kd and Kp, with di�erent methods developed for how the tuning should be performed (e.g. Ziegler Nichols, 'Guess and Check'). The di�erent methods of control shows us that even though only some aspects of reality are taken into consideration when modelling a physical system, we are able to compensate for the di�erences between the model and the reality by introducing a controlling agent, getting closer to having a physical system in reality behaving according to a selected model. 2.6 Virtualisation and Simulation tools The following simulation has been made using Matlab Simulink tool to represent a point mass trajectory system that can be used to simulate a step response of a robot At this basic form, the simulation consists of a trajectory formula that is used as a reference for a PID controller, a constant external force and a set of integrators. In the current model con�guration, changes to the robot trajectory due to external forces can be examined and accounted for by tuning the gains on the PID controller to the desired behaviour. Note that this model is probably a poor representation of reality as it does not take into consideration other variables such as friction forces but is still a valid simulation as long as the limitations are known and accounted for. Figure 7: Step response until external forces 10 Figure 8: Model vs Simulated position and velocity 2.7 Challenges faced creating simulations The above mentioned attributes need to be taken into consideration when creating a simulation model. The challenges that are faced range from deviations in sensor-based control approaches to the friction that cannot be helped in a real-world environment. The relation between reality and model grows ever so closer with growing technology but one cannot help factors such as dust on the oor, which will not be on a simulation. The main challenge is developing new systems which have not currently been implemented into established technology. Another challenge arises when models are scaled up to full size. On a model, it can be scaled down to 1/1000 of a real world size and it might work e�ciently and perfectly on the model. But once scaled to the correct useful size, it might not work as �rst intended and be wasteful instead. This is known as the scalability challenge. Finally, the biggest challenge that can be faced when creating a model is knowing how to use the program to the best of your needs. Not every program will work, and many will not have the components, nor the know-how to create what is simple on paper, into a model on the computer. This is a hard challenge to overcome as it is more trial and error with multiple programs. In our group, we were all used to di�erent modelling programs and had to compromise with Simulink which had not been used by any before. 3 Lean Management Systems Lean Manufacturing �rst started when the Japanese car manufacturer Toyota reinvented their factories after WWII. As the Japanese could not a�ord the expensive machinery and big inventories needed for traditional mass production. They instead started working on a new way of manufacturing, since resources were scarce and their workers were well unionised, Toyota needed a much more e�cient way of production. Where traditional mass production would create gigantic batches of completely identical cars, with of often quite poor quality, Toyota instead focused on making smaller batches of cars, but with a very high quality. This led to the Toyota Production System (TPS) which is known as being the �rst Lean Production system. \The Lean Producer ... combines the advantages of craft and mass production, while avoiding the high cost of the former and the rigidity of the latter. (...) Lean Production is \Lean" because it uses less of everything compared with mass production - half the human e�ort in the factory, half the manufacturing 11 space, half the investment in tools, half the engineering hours to develop a new product in half the time." [12] 3.1 Eliminating Waste The key focus in Lean Production is the elimination of waste. By �nding as many sources of waste as possible and eliminating these, all unnecessary processes of the production are removed. This leaves only the value adding processes, the processes that actually add value to the �nal product. In the eyes of Lean, waste is not just a matter of wasted material, but waste of any kind, such as over- production, big inventories, waiting, moving, transportation and defective products. By removing these kinds of waste a manufacturer will be left with a much more streamlined production, where a product moves through a production line much faster and with a much higher quality. It will also allow the workers to produce more in less time, as a worker in the Lean system does not waste time between work processes. 3.2 Kaizen One of the most important tools in Lean Production is kaizen, it is the part of Lean which refers to a continuous improvement of processes (CIP). This basically means to �gure out ALL the small mistakes, wastes and errors that are causing delays to the production or reduces the output of a production. \Ev- eryday improvement - everybody improvement - everywhere improvement." (Imai) In real life this means everyone gets to be heard concerning errors at their workspace. If any errors are found, it is reported to nearest leader in the management chain. Afterwards the leader will come by the problem and inspect it. This means the processing of errors is dealt by in the most e�cient way, directly at the root of the problems. This method of eliminating errors also gives a great responsibility to all of the employees. Instead of being ashamed of their mistakes they can actively take part in optimising their workstation and thereby contribute to the entire process. This system is called employee suggestion system. This is very helpful for the elimination of errors, as the people who are facing them will usually have hands on experience. The responsibility comes along with the dealing of kaizen and doing it right. In relation to that kind of management there is a method called shop oor management, which works by having the employees meeting up for 20 minutes or so everyday. In these meetings a representative (leaders as well as workers) from di�erent departments of the production discusses their di�culties at their departments. These meetings makes sure that everyone is updated about the greater good of the production and, more importantly, the knowhow about the di�culties they are facing in a broader scope. That is why every individual has the opportunity of taking responsibility in contributing to the elimination of errors. So in kaizen there is fast communication between every step of the management chain. Most important is the fact that everyone is able to pass on their hands on experience of problems. This why the model of Lean �ts so well into reality. Several of experts in their own small �eld who are all contributing with their di�erent aspects and therefore the model are able to take a lot of variables into consideration. The inclusion of the workers in the kaizen method also gives the worker a higher sense of purpose, as the worker can directly impact the work processes of that worker's workstation. 12 Kaizen is often used implicitly, meaning that a worker will incrementally change his or her work pro- cesses for the better. This change could be as simple as a worker moving a trash bin in order not to walk as far during his or her work process. In order to make kaizen work more e�ciently, however, it should be used in an explicit manner. In order to achieve this the PDCA (Plan Do Check Act) cycle can be used. The PDCA cycle works by continuously improving the standards of the work processes. This is done by making every incremental improvement the new standard. To make the process of incremental change standardised, the PDCA cycle uses simple steps in order to quickly �nd more e�cient ways of doing di�erent work processes. To improve a work process one must �rst look at the current condition of this process, how is it currently done, by whom, and how fast? One then describes the desired change to the process. Then the change is implemented and tested for a set amount of time, for example a week or 30 days, or even more. During this test period data from the work process is collected. At the end of the test period the change is either accepted, if successful, or rejected, if not. In this way the methods and tools of Lean Production can be used both implicitly, with loose or no guidelines, and explicitly, with standardised rules and guidelines to follow. Using a model or method explicitly is to be preferred as using methods this way, ensures that all user of the method or tool use it in the same and correct way. Figure 9: Comparison to Mass production [13] The underlying values of kaizen are quite opposite of the view of the worker in traditional mass pro- duction. Where Lean Production utilises the worker's hands on experience and allows for the worker to impact the work processes, mass production instead expects its workers to do their job, as they are told to do it. This means that a worker in a mass production facility has much less joy in his or her work. Mass production also uses huge inventories, as these are seen as a bu�er between the production and the market, though the production would stop for a while, products can still be sold and shipped as a large 13 inventory of �nished products is kept. Lean Production, on the other hand, does not use large inventories, as these are seen as a form of waste. An inventory needs maintenance and supervision. Lean will instead let the product be produced when an order for said product is made. This means that the Lean approach does not push its already �nished products on the market, as mass production would, but instead lets the demand of the market pull the product from the manufacturer. These di�erences make Lean Production a more innovative way of production as change is welcomed, since it is seen as the way to a more e�cient production. \Perhaps the most striking di�erence between mass production and Lean Production lies in their ultimate objectives. Mass-producers set a limited goal for themselves - `good enough', which translates into an acceptable number of defects, a maximum acceptable level of inventories, a narrow range of standardised products. (...) Lean producers, on the other hand, set their sights explicitly on perfection: continually declining costs, zero defects, zero inventories, and endless product variety." 4 Organisational Theories and Modelling Organisations are socially structured, goal-directed, boundary maintaining, hierarchically di�erenti- ated, open systems of human activity, socially constructed indicates that physical, living, human beings interact in shaping and reshaping the system. Goal directed means that, each component of the organisa- tion has a speci�c activity that contributes to the focal goal of either organisational survival or a speci�c achievement. Boundary maintaining refers to the organisation's routine of distinguishing between members and non-members and hierarchically di�erentiated means one or more hierarchical components have the authority to assign activities to one or more hierarchically subordinate components.[14] \Organisation design is the deliberate process of con�guring structures, processes, reward systems, and people practices to create an e�ective organisation capable of achieving the business strategy." [15] Organisational Models serve as a decision-making framework for organisation design. Our group has se- lected to focus on the Star model which is a congruence type of model in order to provide understanding of organisational design and analysis. The STAR model was the focal top choice. The primary reason being that when comparing Star Model to other examples we can notice that there is a great impetus on a process and rewards. This model gives us equal relevance for all factors which means that it secures exibility while being relevant to all types of organisations; structure is not overemphasised; people, rewards and processes can become more important depending on the need. Another important agent which is evident about this model is its adjustability. The strategy implemented is di�erent for di�erent types of organisations. There is no one exact framework which each company needs to �t in. Therefore, this example provides us a role model according to which organisations should be designed. A signi�cant part for this model are alignments between all factors. This model ensures equal interaction within all organisation so that it suppose to work as one body. Star Model has also an ability to overcome negatives of any structural design. In every organisational structure you will meet positives and negatives but as far as management can recognise the negatives, other actors can start working upon up-turning negatives into positives. Models also provide a common platform for decision making in organisations. People in the organisation can debate and provide explanations behind certain decisions made. The advantages of following organisa- tional models includes forcing of key decisions to be based on the long term business strategy and provision 14 of reasons behind the considerations of particular choices. The Star Model portrays a cluster of strategies which organisations need to execute. A strategy implies the set of capabilities in which organisations must excel in order to achieve their strategic goals. Alignment formed the core and fundamentals of the Star model. This can only be achieved when each and every component of the model works together cohesively in order to support the strategies. The goals of the organisations are achieved when the structure, processes, people and rewards are all running in sync. The result of the working of the model in the present day will in uence the plans for the future. Equal importance is also given to the concept of re-alignment among the components of an organisation as the circumstances in an organisation change. The con�gurations of resources and people etc are handy and helpful in order to bring in stability and e�ciency to an organisation but on the other hand can be detri- mental if an organisation is seeking change and exibility in order to recognise its opportunities and threats. Alignment is hence best used as a continuous process which is to be executed (like changing business strategies). The concept is highly dependent on one of the seven Organisational theories called the Con- tingency and Congruence theory. Donaldson [16] states it to be that Organisational e�ectiveness could be achieved by �tting in the organisational features like the structure, strategy to contingencies that re ect the situation of the organisation. As per de�nition \A contingency is any variable that moderates the e�ect of an organisational feature on organisational performance." The theory primarily exists to align organisational features such as the structure, people and strategic decisions made. The contingency theory has further been extended by the economics based complementary systems theory. This theory explains that organisational design choices work coherently. Figure 10: The Star Model [15] 15 4.1 Strategy It is an organisation's formula for success and is implemented in a certain direction in order to achieve the main/sub goals. It is primarily implemented in order to gain a competitive advantage over other competitors. The leadership of the organisation usually are the ones to decide and implement the strate- gies based on the external factors a�ecting the organisation such as customers, technologies, competitors. \Strategy is concerned with the deployment of resources, tactics (or execution) is concerned with their employment." 4.2 People People are at the centre of everything. It is them who formulate strategies and who design and make structures and simultaneously operate systems and processes. People develop and implement the emerging technologies. They are also the ones who bear the identity of the organisation as well as de�ne the particular organisation's culture. As stated famously, "Organisations don't do anything, people do." The skills or competencies people bring to their work are an important factor and so are the values and beliefs they hold. People belonging to various demographics adds value to the resource pool of an organisation. Complex organisations have people at all levels with a particular set of competencies who are able to think and make decisions, interact among each other and participate as teams. 4.3 Structure Structure is devised by the people and management of the organisation in order to achieve the target set by the various strategies. Structures could be created to monitor work ow, run businesses or manage projects. Internal structures for projects tend to be temporary while the core structures of the organisation is permanent. Keeping the di�erent structures independent allows for greater exibility in navigating the direction of the organisation. An organisation?s structure determines where the authority and administrative power is located. 4.4 Processes One of the key organisational challenges is to integrate activities and bridging internal boundaries. Processes play a key role in supporting this integration mechanism. The work of an organisation is accom- plished by people, by machines and by combinations of people and machines. Processes knit organisational boundaries together and determine how well organisations work as a single unit. 4.5 Culture The culture aspect is explicitly not a part of the STAR model since it cannot be designed. However, it is the identity of the people of an organisation. People bring more than themselves and their skills to work; they also bring their attitudes, their values and their belief systems. These interact with, are merged with, are modi�ed by and incorporate the attitudes, values and beliefs that other people bring with them. These interactions occur in the course of working together for formal and informal business, organisational and personal purposes. Emerging from all this is a set of behaviour patterns that is often described as \the way things work around here." 16 4.6 Rewards Rewards are used as incentives to motivate employees to work in an aligned manner. Rewards to the employees who add value to the organisation could be in the form of salaries, bonuses, bene�ts and recognition etc. Rewards are incentives for employees to achieve the organisation's goals and successfully executing the organisation's strategies. [15] 4.7 Technology \The work of the organisation depends not just on the skills and competencies of its people but also on the technologies they employ. Some of the more important skills and competencies tie very directly to the technologies being employed. Companies manufacturing and or selling sophisticated electronics equipment employ various technologies related to electronics". [17] They also employ people like engineers as they are the ones that possess skills and competencies in relation to electronics. 4.8 Problems Organisational Actors Face Organisational actors are key for any model devised. [18] There are numerous amount of problems and complexities that organisational actors can face. In order to have a better overview and to better under- stand these problems, the Adaptive Cycle was developed. According to strategic-choice approach to the study of organisations the model works consistently as well as ideas made by several theorists (Chandler; Drucker; Thompson etc.) are broadened by this cycle. Supporters of strategic-choice approach argue that organisation is only dependent from environmental conditions and from managers decisions. Looking from another perspective, although those factors are signi�cant for the functioning of an organisation, they still can be divided into three main problems of organisational adaptation: the entrepreneurial problem, the engineering problem and the administrative problem. The entrepreneurial problems occurs among plenty of organisations. It is noticeable almost in every new or rapidly growing organisation and among those which walk away from a major crisis. At �rst case it is signi�cant to have an entrepreneurial insight well developed into a well de�ned domain where you exactly know what is your speci�c service or good, a target market or market segment. However, in running an organisation the problem is more complex. The organisation has already determined all their approaches, they know which problems should be faced. Here is the most important issue as the organisation has to �nd a solution for those problems which is always linked with changes in organisation. In both cases man- agement has a last word while taking decision about the solution for entrepreneurial problems. They mark a certain domain at which the organisation would like to work and then they are going to start achiev- ing their target. Commonly, engineering comes into action now, but the role of entrepreneurial activities cannot be forgotten in further stages. This function still stays among top-management responsibilities, however less meaningful managers take care about new opportunities and ideas at an early stage. Then we could meet the engineering problem where you need to develop a system which will be somehow an answer for entrepreneurial problem which was de�ned and identi�ed by the management. Management is responsible for choosing the best and most suitable technology for manufacturing goods or distributing services. They also need to create a new communication chain and to assure control of linkages to secure appropriate operation of the technology. Once the problem is solved and new solutions are being applied, primal version of administrative system is also being implemented at that point. Next issue is that you 17 never can be sure that the con�guration of the organisation will not change after the engineering problem is solved. Current form of organisational structure is dependent from action which takes place in during administrative phase while management tries to set and crystallise relations with environment and to set rules and laws for controlling internal processes and operations. The last problem mentioned is the administrative problem which is almost all about reducing uncer- tainty in an organisational system or when it comes to organisational models it is to solidify those solutions and actions which solved previous problems in the entrepreneurial and engineering stage. Expediting of the systems already developed is not the main issue under administrative problem. The more important and complex task is to bring in processes which will aid organisation to work constantly in an innovative manner. The essential part is to keep the administrative system as a half in rationalisation function and half in articulation function. A perfect administrative system should be both lagging and leading variable in the process of adaptation. A lagging variable means that important decisions which were taken in previous phases need to be modernised by development of new methods and structures. A change in the organisation's processes could be a solution. Leading variable means that the administrative system needs to care about organisation's future to adjust to methods how innovative actions happen. [19] 4.9 Mintzberg's Elements and how organisation models compare to reality In addition to the STAR model Mintzberg's �ve con�gurations based on elements of organisation structuring could also be deemed as models used for organisational analysis. Simple Structure, Professional Bureaucracy, Machine Bureaucracy, Divisionalised Form and Adhocracy hold their unique characteristics as models and are reliant on one or another of the �ve coordinating mechanisms. Primarily, all organisations prefer logically clustered and arranged elements in a particular con�guration in order to attain harmony and consistency within their processes. However, this is where the di�erences between a model and real life organisation can be observed. Organisations in real life will eventually to a certain extent be classi�ed under hybrid models and con�gurations owing to external in uences. The three perspectives of looking at an organisation also play a role here. [20] 5 Discussion The concluding section of the discussion will entail a detailed scrutiny of each domain mentioned above and how they relate to reality and models. This should conclusively discuss how each domain is similar and di�erent. Along with this, simulation, qualitative and quantitative experiments will be explained in relation to reality and models. The last part will entail why validity and reliability are di�erent and how they a�ect the domains. 5.1 The 3 domains 5.1.1 Similarities between the 3 domains We identi�ed the three course domains to share a similar process of optimisation. All three domains border a set of theories , which are used to perform a well established observation of a process and a makes a basis for introducing a model. According to the examination analysis and the domain in which it exists, an ideal model is applied . Becoming a reference point for all future Improvement. As explained previously , in the process of modelling, not all factors are accounted for, which may result in a di�erence between the real and the expected (modelled) process outcome. The di�erences can be 18 accounted for by introducing a control loop (Feed-forward / Feedback) which provide the adjustability and continuous improvement needed for successful model integration. The following examples show how di�erent optimisation requirements fall under the three domains , and demonstrates the similarity of the processes within the three domains: Figure 11: Optimisation - Applying theory to reality If a robot's step movement needs to be improved, the modelling and control domain glass is put on as it contains physical and mathematical theories which can be used to generate a di�erential equation and the physical constraints needed to create simulations and to tune to the desired behaviour by introduction of a PID controller and comparing to the reference equation. If an organisation's performance needs to be improved , the organisation theories glass is put on as it contains theories which examine speci�c aspects of an organisation and models which have been proven to increase the performance of a desired aspect, the improvement can be then be compared and tuned as needed. If waste should be reduced in a shop- oor, the lean manufacturing glass will be put on as it contains theories which examine the di�erent sources of waste and o�ers lean models that contribute to waste reduction such as Kan-ban, improvement is then maintained by receiving continuous feedback from the employees. 19 5.2 100% The three domains all aim to have 100% output while keeping the cost close to nothing. This is easier to simulate reality as explained throughout the report. The 100% e�ciency is where every company would like to be but it is impossible. To reach this, there must be absolutely no waste, no breaks, no sick days, which in a real life situation, cannot happen every day of the year. This means a company aims to be as close to this. Kinematics, lean and organisation theories, once combined, can improve a companies e�ciency, validity and reliability (section 5.4) tenfold. 5.3 Qualitative and Quantitative Experiments Simulation, qualitative, and quantitative experiments [21] are all necessary to perform more accurate experiments and create better models. Qualitative research provides an insight to opinions of others around the area of research. Typically, the sample sizes are not big, and interviews are done individually, instead of in groups. Quantitate research is number based and often a lot more accurate and solid data. An advantage of only using quantitative testing is that it is easy to manipulate mathematically and re ects the researches point of view. Qualitative and quantitative experiments have both advantages and disadvantages. While qualitative experiments deals with a more descriptive data, and quantitate uses hard data and numbers to perform experiments. Using just one of these methods may prove to have aws due to the limitations in qualitative or quantitative experiments, and therefore a combination of both have to be utilised in order to deliver the best result possible. 5.4 Validity and Reliability The two words, reliability and validity form the fundamental base of scienti�c research and judged to be as the minimum criteria and scienti�c proof for any piece of research. The idea of reliability could be relevant in all the three domains of Kinematics, Lean Methodology and Organisational theory. As the course entirely is based on converting and comparing and re ecting on modelled theories into real life. Validity and reliability can be observed in all the three domains. An example in the Organisational Theory domain could be the implementation of the STAR model into reality. Due to its exible nature the STAR model should be able to be implemented into designing any type of organisation. These two key-words are familiar in experimental design and results. However, the STAR model could also prove to be valid by providing the expected results when the management follows the required guidelines. On the other hand, the validity factor is highly dependant on how accurately the organisational design was planned and executed. In comparison to reliability though the model might fall short in producing results as it would not guarantee the exact expected outcomes for di�erent organisations. In the domain of Kinematics these two words are much more relevant. The simulation of a body in free space is akin to conducting an experiment. Therefore, both accurate and precise results are expected when describing a point object in cartesian and polar coordinates. The simulations should also be able to be re-creatable and repeatable and should obtain the same or similar results with minor room for error when exactly performed by other researchers as it is explaining the physics and dynamics of a system. In the lean theory domain similar examples could be observed and evaluated where the concept is perfect as a model but cannot be reliable when implemented in di�erent circumstances unless each element of the company perform and behave in an aligned manner. 20 References [1] What is theory? Research Methods for Business Students, Sixth Edition: Saunders Mark, 2012, p45 - p47, Chapter 2 [2] De�ning theory http://www.businessdictionary.com/de�nition/theory.html [3] Box Box, G. E. P. (1976) Science and statistics, Journal of the American Statistical Association 71, 791- 799. [4] Models Representing Reality Philosophy of Science, RonaldN.Giere, Vol. 71, No. 5, Proceedings of the 2002 Biennial Meeting of The Philosophy of Science Association [5] Model based theories Robust Autonomous Guidence, An internal Model Approach, Isidori, A; Marconi, L; Serrani, A, 2003, p1 - p4 [6] Walton's theory Adam Toon, Synthese, Vol. 172, No. 2 (Jan., 2010), pp. 301-315 [7] How to de�ne what is Reality and what is not http://www.world-of-lucid-dreaming.com/what-is-reality.html [8] Reality White et al. 1982:14-15 [9] Simulation Meaningful Learning with Technology, Fourth Edition, Cream101, 2013, p132 [10] PID de�ning `PID without a PhD', Wescott Tim, October 2000, EE Times-India [11] Control loops Control Loop Foundation - Batch and Continuous Processes, Terrence Blevins, 2013, p22 - p26 [12] Lean Production De�nition Womack Jones Roos, 1990, p13f. [13] Wheel of improvement https://en.wikipedia.org/wiki/PDCA [14] Organisational theory APA handbook of Industrial and Organizational Psychology (Sheldon Zedeck) , Chapter 5: Organiza- tions: Theory, Design and Future, (George P. Huber) [15] Star Model Designing Your Organisation, Amy Kates, 2014, p1 - p26 21 [16] Donalson Lex Donaldson [17] Nickols The Organizational Analysis Model, Fred Nickols, 2012 [18] Ideas to Reality in organisations Wells, Susan J. \From Ideas to Results" HR Magazine. Febr., 2005 [19] Raymond Organizational Strategy, Structure, and Process, RAYMOND E. MILES (pp 546-562) [20] Mintzberg Article Henry Mintzberg, Management Science, Vol. 26, No. 3 (Mar., 1980), pp. 322-341 [21] Qualitative vs Quantitative Data Algebra Lesson Page, Donna Roberts, 2012