Module 5: Modelling Complexity

Module 5: Modelling Complexity

“Modelling: why and how? … Simulating the energy transition model … Gaming simulation … Discrete-event modelling & simulation … System Dynamics modelling & simulation … Agent-based modelling & simulation”
(Source URL)

Summaries

  • Module 5: Modelling Complexity > 5.3 Gaming simulation > Web lecture: Gaming simulation, Part 1
  • Module 5: Modelling Complexity > 5.3 Gaming simulation > Web lecture: Gaming simulation, Part 2
  • Module 5: Modelling Complexity > 5.3 Gaming simulation > Web lecture: Gaming simulation, Part 3
  • Module 5: Modelling Complexity > 5.4 Discrete-event modelling & simulation > Web lecture: Discrete-event modelling & simulation, Part 1
  • Module 5: Modelling Complexity > 5.4 Discrete-event modelling & simulation > Web lecture: Discrete-event modelling & simulation, Part 2
  • Module 5: Modelling Complexity > 5.5 System Dynamics modelling & simulation > Web lecture: System Dynamics modelling & simulation, Part 1
  • Module 5: Modelling Complexity > 5.5 System Dynamics modelling & simulation > Web lecture: System Dynamics modelling & simulation, Part 2
  • Module 5: Modelling Complexity > 5.6 Agent-based modelling & simulation > Web lecture: Agent-based modelling & simulation, Part 1
  • Module 5: Modelling Complexity > 5.6 Agent-based modelling & simulation > Web lecture: Agent-based modelling & simulation, Part 2

Module 5: Modelling Complexity > 5.3 Gaming simulation > Web lecture: Gaming simulation, Part 1

  • Hi, my name is Sebastiaan Meijer and today I would like to tell you something about a topic called: gaming simulation.
  • What is it? It’s about system thinking and the actors in the systems.
  • Then I’ll tell you exactly what game simulation is and I will illustrate this with two examples.
  • So if you think about systems – and if you think about social-technical systems especially- then there are always people involved.
  • These people have a choice in the way they do governance, management and operations of a particular system.
  • For instance , if you talk about the transportation system, then the individual travelers – of course- have a choice on what they do every day.
  • The second view is that of complex adaptive systems.
  • They have not only all kinds of unpredictable behavior, but also they have all kinds of relations and intertwining.
  • All these unpredictable components make that the performance of such a complex adaptive system can better be considered as emergent behavior than as something that is very predictable or a result of something.
  • That there are people that analyze all kinds of system dynamics.
  • What if we think of a complex adaptive system, there are all kinds of actors that have relations to each other, This is what we could call: a social network.
  • If we consider technical components, take for example again the transportation system, then we have trains, we have cars, we have roads.
  • Which also might comprise of institutions or companies connected to this, jointly form a complex adaptive system.
  • You can see in the table that they have all kinds of V-models, for instance in systems engineering.
  • Does my division perform well, does the system perform well? And to do so – you see the logic – they try to find acceptable and reasonable solutions.
  • They often use all kind of mixed tools: project and process management tools for instance.
  • If we have all these types of actors together, we need to find solutions in which they can work together on formulating new states of a particular system.
  • In a gaming simulation session, so that is the play of the game, we mimic the best we can behavior of real world system.
  • No. We put real people into the role of the decision maker and this we combine with computerized simulation models very often.
  • It goes from analog, multi actors, stakeholder exercises, to complex 3-D multi environments and games in which participants really act together with computers to make simulations.
  • In the 1970s, you see that – led by amongst others Duke and Gibbs in the US, and the UK stream by Checkland and Scholes – testing of complex systems becomes important.
  • An important one to mention is Martin Shubik who wrote a book called the ‘uses and methods of gaming’ already published in 1975.
  • In the 1980-ties, you see computer gaming coming up and at the end of the 90-ties this gets into serious applications.
  • You can nowadays quickly divide a class into different generations by playing the sounds of the different computer games at that era.
  • Some generations will know paratrooper, some will know Mario Bros and I’m absolutely sure that the younger generations will be completely steered by the sounds of the games that are coming out right now.
  • What you see is the integration of-on one hand this policy making and on the other hand the computer gaming.
  • I put in the middle of this figure the session because for us the play of the game is the most important thing.
  • The input into such a session is a game, – a gaming simulation design – , that specifies the role, the rules, objectives and constraints that you need to be able to play a particular system.
  • It depends a lot whether I play a game with you in a classroom setting or in your daily practice.
  • Or if the consequences of winning game are passing the course or just having a nice time.
  • These participants become input into session, you play with them and afterwards they have a particular experience that you can change into learning or into change behavior.
  • Getting data out of the play, out of what people do, is a fairly new scientific method in approaching games.
  • We evaluate the change of organizations or we evaluate the learning over the game.
  • Whereas – from left to right – we are in an analytical science fashion and in the analytical science fashion, this is the red feedback loop at the bottom, we try to say something about the real world by using the data generated in the game session.

Module 5: Modelling Complexity > 5.3 Gaming simulation > Web lecture: Gaming simulation, Part 2

  • I would like to go further in the lecture on gaming simulation by showing you 2 examples.
  • In the Netherlands, we have a separation between the company that operates trains -there are actually multiple companies- and the public agency that operates the rails and does the traffic control.
  • There are engineers that really know how the system work and they test their potential solutions in computer simulations.
  • Will it work in practice? We know in many railway systems that robustness and resilience is already under pressure.
  • A nice phrase here is that the difference between theory and practice only exists in practice.
  • If you start doing it with them – in a game – then actually you get a lot better feedback.
  • So the challenge that we have here is: we want to have a 100% extra trains in 2020.
  • The goal was to have 50% extra trains in the Western regions of the country by 2012.
  • With a metro you don’t know when the actual metro comes you just go to the station you pick the first train.
  • That would be great to do with the train system as well.
  • So can we do that smarter? The project ‘every ten minutes a train’ was a real world test done in September 2010 between Amsterdam and Eindhoven in what we call a 6-6-2 pattern.
  • This means 6 long-haul trains, 6 short-haul trains and 2 cargo trains.
  • The preparations started in spring of 2010 and the question was: what if a disruption occurs? How can we handle this? There were 2 types – the traditional type and the new type – defined by staff experts.
  • NS is the major railway train operator and together with Prorail and us, as facilitators, we held this gaming session.
  • Very simple models, but with real numbers to it and they represent trains.
  • You also recognize that there are all kinds of things plugged into these scour sponges and these represents train drivers and heads of cabin.
  • This manual simulator – as you could call it – was controlled by these people sitting in control rooms and being connected by 2 radios and video streams.
  • We modelled trains by scour sponges and combinations of wagons of a particular type.
  • There was one train driver and one head of cabin required, and those were flags with the real numbers.
  • We also have rootsetting, again on real map, and a timetable with real numbers even though it was the future timetable with 50% more trains.
  • Now you could think: ‘okay if it’s so simple to simulate something what type of models do we actually need to use’? I often make a distinction between what I call an iconic representation, and if you see this icon on the screen anywhere on the airport you know where to go to go to the train.
  • Some things you need to make as if real, and these are crucial elements the people really need to pay attention to.
  • So here I have 3 pictures of a train and they mean something entirely different.
  • For computer simulation doesn’t make sense but for gaming simulation – where we have people – it makes a lot of sense to do so.
  • It is is about the relation between urban development and the development of the transport system.
  • It has been developed for an agency called Delta metropool foundation and it is about connecting the urban inner city areas through metro-like trains again.
  • We build a computer based game here to be played with domain specialists of all kinds of backgrounds.
  • We could give them fictitious, or even real, roles in this game.
  • With this, we built a computer game – and you can see here at the bottom of the screen that we have groups playing that game.
  • If you only build houses in one place and offices in another, you are going to overload the transportation system.
  • Behind a purely qualitative interface to make sure that every actor -even if he’s not from an engineering type of logic- can play this particular game.
  • The learning outcomes are that the complexity and the dependency in urban development between transportation and the success of cities is very important.
  • I believe that gaming, computer simulation and big data – so trying to capture the entire world by all the digital traces that we leave behind – this going to integrate.
  • This will lead to multi-scale modeling and data-driven gaming more and more.
  • In the visualization and the interactivity computing domain, you see that the technology of using 4k projectors and the extremely large datasets becomes useful for gaming users for real time interactive use.
  • The politicians know that they need the knowledge of how the system actually works – same with the managers – to make good decisions and this is where we will use joint fact finding.
  • You could visualize this as if you have game simulation.
  • Not just politicians – or the high level decision makers – but also the operators and the engineers.
  • Gaming simulation is one way to make a contribution here.

Module 5: Modelling Complexity > 5.3 Gaming simulation > Web lecture: Gaming simulation, Part 3

  • Hi. Welcome back in the lectures about gaming simulation.
  • Today I would like to show you some examples of the 4 uses of gaming.
  • You could say that there is intervention where group learning, policy intervention and the theory by Duke and Geurts on the 5 C’s of gaming is important.
  • Last but certainly not least: a game for design.
  • In 2011 this got halted and then we were asked to make a game for all the vice commanders of police in the Netherlands to play ‘the future organization’ in which intelligence based policing was implemented.
  • Do we actually want this? For teaching, I have an example from an oil company where we build a game that trains participants to recognize dangerous tasks.
  • So it is a 3D game built here in a house where we mimic an existing oil drilling site.
  • This is a typical gas well, and there at the back you see that there is a crane and a crew that has to do a particular task.
  • This task has to be done in a safe way and therefore you need to interact with all the people in the game to make sure that they do their task in the correct way.
  • That is a good thing about these types of training games: it can be a catastrophe but then you just press replay and there you start again.
  • The third use is games as an empirical test environment.
  • Can we test hypotheses? Can we test our ideas about designs and how they will behave with the people in the loop? Here I have an example from Sweden.
  • What you see here is all kinds of people from different backgrounds where the road administrations, and these are people that really have this role.
  • They set the criteria in the game for what they actually want.
  • We played this game during the day, in which there were also seminars.
  • What you see here is that people actually engaged so much in their roles – because they are real contractors – that even at a very cosy diner at a castle they are really putting in their roles and putting their bits and trying to win the game.
  • The Social space which involves: the people that use it, the knowledge, the community around a particular problem and product.
  • The thing is, that if you want to do design of these complex systems, that you really need to connect this P, S and I space together in the design.
  • Here we use games as a way to go back and forth between the different spaces.
  • To characterize a problem in these different domains, we will bind together the engineers that work in the P space, managers that work in the S space and the politicians that work in the I space Let me give you an example again from Prorail.
  • Here again we use fairly simple tools that real people, real experts and real people as experts trying to simulate the effects of new designs.
  • So together with the people that know how it looks in reality we designed a test and came up with better solutions.
  • Here we really proved that gaming is a way to go back and forth between changes in these different spaces.
  • As was the word serious games about 5 sometimes 8 years ago.
  • What does gamification actually mean? We try to influence the behavior of people in a positive way by making systems change into a game.
  • It is not simulation, but really using game elements to do something in the real world.
  • If you take this a few steps further, you can make many things into a game.
  • Now in the era of social media you see that platforms like Facebook, like Myspace and some other platforms, are actually trying to make their platform the key place for these gamification things to happen.

Module 5: Modelling Complexity > 5.4 Discrete-event modelling & simulation > Web lecture: Discrete-event modelling & simulation, Part 1

  • Topics I’m going to focus on are: What is a discrete event simulation? Where does it fit historically? How does it differ from other types of simulation, such as agent based simulation and continuous simulation? What are the steps in a simulation study and what are the important aspects of simulation for infrastructure studies? According to Shannon, 1975, simulation is a process of designing a model of a concrete system and conducting experiments with this model in order to understand the behavior of the system and/or to evaluate various strategies for the operation of the system.
  • We change parameters to the simulation model and look at the effect on the output of the simulation.
  • In continuous models states, and the state variables are a continuous function of time.
  • It also makes the discrete event simulation models pretty fast because the state only has to be changed and these point in time.
  • Models for the current situation and models for potential future situations.
  • The models have to be identified based on states and information we can get from the infrastructure system itself.
  • We build a model, diagnose the problem, test whether the model is corresponding to the real system, build alternative models to evaluate them for solutions.
  • This means that we build a model, we go to specification, we go to data collection, we find in the data collection that maybe we need to make some changes and then we go back to one of the earlier phases to adapt the model.
  • An incremental model is lately seen as the best way to build discrete-event simulation models.
  • We start with a tiny model that models the core of the problem or of the infrastructure we are looking at.
  • Step by step, we extend to model in a number of steps where we go through all the phases of the model cycle again.
  • Starting small and extending it step by step to a bigger model.
  • Especially for larger infrastructure models this is an excellent way to do it because we don’t have to make the model all at once leading to all kinds of bug finding and mistakes.
  • It means that we find the system boundaries and we determine what goes into the system, when it goes into the system and what belongs to the system and what belongs to the environment? We also establish the components, the objects the entities that are in the model and the processes that we see in the model.
  • This is, for example, a model of parts of the infrastructure of a large airport and with this we can look how we built up the model from smaller pieces.
  • The specification, the next step in the model phase, is aimed at leading to a working model that one can experiment with.
  • Reduction means that we leave out everything that’s not needed to complete the model and to run a successful simulation study.
  • We specify the model, we gather data and we build the model.
  • An example how to build a model is shown here in one of the simulation modelling packages called SIMIO that’s available and often used in university surroundings.
  • With these building blocks, we very quickly built up the model connected by, in this case path, and we have a generation of entities, a server that serve the entities and a sink in which the entities again disappear.
  • The model has both two dimensional and three dimensional representations and one can see that in one minute one builds a queuing model that would normally take, if you would have to program it, a much, much longer time.
  • To give an example, you can see from this very small model that quite a lot of output is generated for later analysis and it’s this analysis that so important also for discrete event simulation because the models are aimed at generating this particular output.
  • What kind of data do we need to fill these kind of models? Especially generators and sinks at the sources and ends of the model are very important.
  • Means that we have to know when entities enter the model and when an entities leave the model.
  • We need to know that our model is a valid representation of reality and there are different ways to test that.
  • Verification looks at the translation of our conceptual model into a simulation model.
  • Validation looks at the relationship between our simulation model and reality.
  • One is to test hypotheses using statistics, where we compare output from our simulation model and output from reality under similar circumstances.
  • In principle every step that we take in the simulation modelling study has to be followed by a validation or a verification step where we test the correctness and the accuracy of the model that we just created and to look whether the model is still really fit for purpose for what we try to accomplish with the model.
  • When the model is considered to be valid and verified, we want to look at experiments of course.
  • We determine the run controlled conditions under which the model is experimented with and thereby we can establish a statistical validity of our model.
  • This is what we do in analysis and diagnosis, where we use the model to calculate results that are statistically valid and from which we can draw conclusions about our infrastructure models.
  • An example is to export model results to, for instance SPSS, or one of the statistical packages and analyze it in detail.
  • If we look at infrastructure simulation, discrete event simulation is very much suited because we can represent model components one to one and, in the form of entities or building blocks in our simulation model.
  • The animation can help in building, debugging and presenting the model.
  • We have also looked at the model cycle, where we focused on incremental building and building blocks.
  • We also will see later, in some of the examples for discrete event simulation in one of the next presentations, is that we focus very much on hierarchy, where we can build up models in a hierarchical fashion and that means that the modeler always keeps overview of the model itself.
  • This means that we have to very carefully look at the boundaries of our system and conceptualization and the reduction our system in the specification to make sure that we have the right input for the model and we can collect the data.

Module 5: Modelling Complexity > 5.4 Discrete-event modelling & simulation > Web lecture: Discrete-event modelling & simulation, Part 2

  • Hi, I am Alexander Verbraeck, professor of systems and simulation and I’m going to look with you to a number of discrete event simulation examples following up on the discrete event simulation module that you have looked at before.
  • Simio, Simulate, Plant Simulation, Enterprise Dynamics, Extend and AnyLogic, and there are many, many others, there are hundreds of different types of simulation software on the market.
  • Some of them are more suitable for some types of infrastructure models and others more suitable for other types of infrastructure models.
  • Building models that can become more and more complicated.
  • All kinds of input and output from and into databases, excel, statistical packages and all kinds of other input and output software Link to optimization software and all kinds of extensions.
  • That means that we can mix discrete event simulation models, continuous simulation models and agent-based simulation models for the behavior of people in our models.
  • What we observed in a number of studies is that every time we had to build a model for solving problems at an airport we had to make a new model.
  • Our challenge in this project was: can we construct a model out of building blocks? Small blocks from which we can build a model bottom-up? Can we thereby tackle all kinds of airport problems in a much more generic way? Our goal was to develop one set of simulation libraries for airport logistics, for airport design and for airport development.
  • When we zoom in on one of these building blocks, for instance on one of the concourses, we see a smaller model; a couple of gates, a couple of walk areas, a couple of conveyor belts that help the passenger to go from or to the gate and from there to other parts of the airport.
  • Then they wait at the gate to get their boarding passes checked an through the bridge they can actually leave the model into the plane.
  • These building blocks are very similar and are re-used in the model.
  • The one thing that we did for this particular model is that we can parameterize the building blocks.
  • One of the interesting things is that we can also animate the model on the highest level.
  • We see it actually getting more and more busy – also at the wait areas – and it’s only a matter of time until the first planes depart and we have actually people going out of the model.
  • One of the most important things is of course the fact that we can also produce output from the model.
  • Here we actually see the model that we’ve just watched in terms of the number of passengers at a certain desk row for checking in.
  • Statistics are one of the main reasons for creating simulation models, and especially discrete event simulation models.
  • We can very quickly model the infrastructure due to the fact that we have high-level building blocks.
  • The models are still complex though and a lot of behavior is hidden.
  • It still is difficult for people to create a good model from this.
  • We really want to compare alternatives in all kinds of different scenarios and the question is: how can you parameterize those scenarios? What you see on the screen here is a very simple model at the top in terms of the weights represented.
  • The model at the bottom shows the background of the particular model, the way it is being built, and we can see that it is extremely complicated.
  • How can we create a flexible set of models for demonstration and teaching of these very, very complicated global infrastructures? This is a view of the way a model looks: We see the fact that on the left side we have initial suppliers, we have on the right hand side all consumers and customers and in the middle the focal company we want to look at.
  • We see a lot of different graphs that we can create with this particular model and they provide a lot of insight into what happens in the different parts of the supply chain.
  • We focused very much on the output here, contrary to the models that you saw in the first case study.
  • In this particular case we used Java libraries to build the models to have full flexibility in the way we could build them.
  • Finally, serious games have been developed with discrete simulation models as the core, which aligns very much with decision making as human decision makers do and they can enter their decisions into the game.
  • In this particular case our challenge was: can we create micro simulation models with a lot of detail that we can use to support decisions 10, 15 or 20 years ahead. Long term policy studies.
  • The question was: are these models fast enough and would they be usable in this particular setting? In this particular case we created the micro level simulation model for barge transportation in the Netherlands usable in policy setting that can be used 10 to 20 years ahead. The model itself is based on a geographical information system and uses a lot of different databases to support the model background.
  • When we slow down the simulation time a little bit, you can actually see individual ships sailing and the moment I click on one, you can see the information about the ship, the number of containers it contains, where it’s coming from and where it’s going to.
  • Each individual containers is modelled for this particular simulation mode, several millions a year! And still the model actually completes a year of simulation in about one minute.
  • The locks and bridges can also be questioned and queried and we can see for instance whether certain types of ships would be able to sail those particular locks and bridges.
  • All these things can be done with one simulation model that we used in quite interactive settings.
  • We really were able to support long term decision making with these particular models.
  • This discrete event simulation formalism took care of creating very fast models.
  • Participants could use the models and were fully engaged in the sessions.

Module 5: Modelling Complexity > 5.5 System Dynamics modelling & simulation > Web lecture: System Dynamics modelling & simulation, Part 1

  • In this weblecture we will look at one of these techniques, the System Dynamics Modeling and Simulation method; what it is used for; how it is used and what is specific about SD? We will also briefly look at two examples.
  • What is System Dynamics? It is a method for modeling and simulating dynamically complex issues or systems characterized by feedback and accumulation effects.
  • What is System Dynamics used for? Quantitative SD modeling is used to model systems and issues and to simulate their behavior.
  • System Dynamics models are also extremely useful for model-based policy analysis and for adaptive policy design.
  • System Dynamics models have influenced and driven energy policies and infrastructures in the US and elsewhere since the 1970s.
  • The same is true for resource dynamics and management from the moment the first “Limits to Growth” study, commissioned by the Club of Rome, was published in 1972.
  • In line with the follow-ups of the `Limits to Growth’ and the growing believe that pollution, more than finite resources, poses severe limits to growth, the System Dynamics field also shifted towards environmental policymaking.
  • System Dynamics has also been used since Jay Forrester’s Urban Dynamics study for studying urban dynamics and policies.
  • Recently, a System Dynamics model of the relationships among many core city systems was developed for the Smarter City project of the City of Portland, Oregon, USA to better understand urban dynamics, and to identify opportunities to become a smarter city.
  • System Dynamics is used for strategic planning and all sorts of business dynamics problems.
  • Think about books like `Business Dynamics’ or Strategic Management Dynamics, System Dynamics models are also used for integrated risk-capability assessment and Critical Infrastructures Protection.

Module 5: Modelling Complexity > 5.5 System Dynamics modelling & simulation > Web lecture: System Dynamics modelling & simulation, Part 2

  • What do System Dynamics models look like? Let’s have a look at the technique.
  • Two types of diagrams are often used in SD. The first type, Stock-Flow diagrams, are mainly used to build simulation models.
  • Apart from stocks and flows, constants and parameters are also included in these models.
  • Auxiliary variables are included to build models that closely correspond to the real world system.
  • If a model is fully specified, it can simulate behavior over time.
  • This type of representation is good for building models and for communicating stock-flow structures, but it is not really appropriate for thinking in terms of feedback loops or communicating important feedback effects.
  • These show causal links between the main variables, the polarity of these causal links, the feedback loops, and the polarity of these feedback loops.
  • A feedback loop consists of two or more causal links between elements that are connected in such a way that if one follows the causality starting at any element in the loop, one eventually returns to the first element.
  • A feedback loop is called a balancing loop if an initial increase in variable A leads after some time to a decrease in A, but also if an initial decrease in A leads to an increase in A. In isolation, such feedback loops generate balancing or goal-seeking behavior.
  • A feedback loop is called a reinforcing loop if an initial increase in variable A leads after some time to an additional increase in A and so on.
  • In isolation, such feedback loops generate reinforcing behavior.
  • Feedback loops are often strongly connected, and their relative strength changes over time.
  • Complex system behaviors often arise due to such shifts in dominance between different feedback loops in the same system.
  • When dealing with feedback loop systems consisting of multiple loops, it is hard to derive the behavior of the system without simulation.
  • If we further transform the Stock-Flow Diagram into a Causal-Loop Diagram, we need to add the link polarities, identify the feedback loops, and derive their loop polarities.
  • In this case we would end up with a reinforcing loop with a delay and a balancing loop.
  • Depending on which loop is dominant and whether dominance shifts, this feedback loop system could generate several modes of behavior, for example exponential growth if the reinforcing loop remains dominant or S-shaped growth if the balancing loop takes over after some time.
  • What’s so specific about System Dynamics? 1) SD models are largely endogenous theories, that is: model boundaries are chosen such that all important feedback loops are within these boundaries 2) SD models are rather aggregated: stock variables are often used to group rather homogenous individuals or items.
  • As already mentioned: SD models are integrated numerically to generate the behavior over time.
  • System Dynamicists are far more interested in improving our understanding and changing faulty mental models, and generating general policy insights than generating predictions.
  • SD models are essentially tools for thought.
  • More, reflection beyond the model is also hugely important in SD. Let me give you two examples now: a very simple example first and then a more advanced.
  • Let’s make a very simple model about the electrification of the European car fleet between the year 2000 and the year 2100.
  • The model displayed here is a bit bigger than the previous model.
  • It consists of a demand sub model, a supply sub model, an extraction infrastructure sub model and a recycling infrastructure sub model.
  • The model was made with mineral/metal scarcity experts and used to develop scarcity scenarios.
  • You are ready to make and use your own SD models and combine SD with other methods.

Module 5: Modelling Complexity > 5.6 Agent-based modelling & simulation > Web lecture: Agent-based modelling & simulation, Part 1

  • This is the next installment of the NGI MOOC and today, we will be talking about Agent Based Modeling, but before we start let’s consider something that you’ll do tonight.
  • So how does one even go about trying to understand all these interconnected systems, that adapt grow and evolve? Well, this is the lecture on Agent Based Modeling, so this is obviously how we are going to approach this.
  • Agent Based Modeling, what is it? And why should you care? So what we will discuss today is the following: We will talk about what an agent is, what agent based modeling is, I will show you a number of examples to help you understand when and how to use agent based modeling and I will try to give you some pointers to help you explore Agent Based Modeling on your own.
  • Well, Agent Based Modeling is a technique that comes from the Complex Adaptive Systems toolkit.
  • ” Now this approach is called Generative Science and this is what is the heart of Agent Based Modeling.
  • How this is done, well we will start by situating an autonomous heterogeneous agent, or entities or things, in some kind of relevant environment in space a network or whatever.
  • How does this work? Well, we sometimes say that an Agent is a thing that does things to other things.
  • You see that at the middle there is an agent, an entity or thing that knows things and does things.
  • Now this thing is situated in a model in some kind of environment, where there are other agents, which sits inside a model that we as a modeler have chosen and drawn boundaries around.
  • Now, looking at these agents, following inputs from others, from the environment and from their own behaviors, they will make decisions and they will perform actions.
  • Now these actions can affect themselves, they can affect the environment around, they can affect other agents.
  • Now there even might be many other agents that are acting, and interacting on their own, without having direct interaction with you, the agent.
  • Now, what are the components of such an Agent based model? Well, first, clearly it is the agents, the entities, the things themselves.
  • The agents themselves, as I said, they are things that do things to other things.
  • This agent can be anything, really anything that you care about and you think is worth exploring.
  • Now the actions, the consequences of the decision making processes, the result of those and Agent will perform or not perform some action.
  • With agents you’ll say it’s your turn to act, the agent will then consider its environment, consider its state, make the computation that it needs to do, and then decides whether to act, and how to act.
  • Now, this action can be based on some input of another agent, on some state, or some kind of internal decision rule.
  • These actions can also affect other agents, they can affect their own rules, their own states, they can affect the environment, and it is often through this indirect interaction throughout the environment that the true complexity of the system arises.
  • Now the environment is a thing the agent is in, and everything that is not the agent but is relevant.
  • Now this environment provides the agent with structure and information.
  • For example if I’m an investor agent trying to decide whether to build a power plant, I might consider the cost of capital, the interest rate.
  • So the environment affects the agent and in some cases the agent may affect it as well.
  • An agent can be on a grid so I can have eight neighbours around me and I only know of those eight, or I could be in a physical space, for example on a GIS map so I can say well this agent is three kilometers away from me or two meters.
  • Now time is a very peculiar thing, excuse me, in Agent Based Modeling, as time takes place in discrete steps, in ticks and between two ticks we assume that everything happens at once.
  • Now it is this parallelism of the real world that we are trying to capture when we build Agent based models.
  • Or the sequence of agent interaction, is very important, because like you can probably imagine if a certain agent always goes first, for example is always allowed to buy first, it will buy the cheapest resources for example and always have an advantage.
  • So we shuffle the agents every time step, and allow them to take turns to form interactions by this we simulate parallel action that we observe in the real word.
  • Now as I opened with this wonderful blue marble image of our planet, what has happened there is individuals arriving at home experiencing the environmental change, they see dark, they see other agents, choosing to turn on the light or not, and based on inputs of others, on their own individual preferences, they decide to turn on the lights.
  • Now, what airbus has done in this case, it has simulated human behavior, individual behavior, in a computer as agents.
  • Now this agents follow rules that are observed when humans are in panic and try to run away and they can simulate time and time over to see how quickly people can evacuate the plane.
  • Now in this case the agents are multiple departments, they all have their own internal logic, their internal goals they are trying to achieve and they are trying to cooperate and make money at the same time and they are experiencing these delays as for example when there are disrupted shipments.
  • So here we have the agents that are energy companies, own multiple facilities that experience changes in prices and changes in electricity demand, and they have to make decisions whether to invest, whether to divest, whether to buy one kind of fuel, or another kind of fuel, interact with each other and the overall emergent pattern will be various levels of CO2 emissions and CO2 reductions and various electricity prices.
  • You can still use an agent based model in a very useful fashion to explore how individual interactions, individual decision-making processes affect an overall pattern.
  • This is power plants making decisions, investing into new facilities, you see that the generation profile is changing, there is more electricity produced from one source than from the other, you’ll see that the CO2 market is changing, you can see all sorts of dynamics happening over time as agents act and react.
  • Now, with this I would like to thank you for your attention and I hope I have given you a basic understanding of what an agent based model is and how you can use it.
  • Netlogo is a free and open source Agent Based Modeling platform that is very well suited for teaching and studying agent based models.
  • The first one, SPM4530, deals with complex systems and a short introduction into agent based modeling and the other course, SPM9555, deals with advanced agent based modeling and it takes you in great detail in how to build complex models.
  • Finally, when you’re ready to dive even deeper there is a book called Agent Based Modeling of Socio-Technical Systems, edited by me, my colleagues Dr. van Dam and Dr. Lukszo and there it is online, you can check it out to dive into this topic much deeper.
  • Now, I wish you a lot of fun and enjoyment studying Agent Based Modelings and good luck.

Module 5: Modelling Complexity > 5.6 Agent-based modelling & simulation > Web lecture: Agent-based modelling & simulation, Part 2

  • Also as a researcher it is fascinating: it’s a huge source of CO2 emissions, has a few of the newest coal power plants, making it a large electricity production center, and it is one of the largest energy consumption areas.
  • Agent-based models are a computerized laboratory, which help to explore what goes on here at the Port of Rotterdam area, specifically in the power sector, because they do not ignore, but rather embrace the complexity of the system! How can we start to understand and shape the system in the direction we want it to go? How can we get this system to change? We are so keen on reducing CO2 emissions – well, the things that need to change are right before our eyes!! As an agent-based modeler, I immediately start to think: well if you want to lower our emissions, we need to change someone’s behavior.
  • So… what are the long-term effects of climate policies? What model can help you to tackle the problem? At least you need the existing power plants in there and the owners operating them.
  • You represent the power plants, model the owners as agents that do the operation: they sell electricity to the market – representing the consumption – and choose when to run which power plants.
  • You make sure that they are smart enough to run the cheapest plants: at lower demand the cheaper plants run and at peak demand also the expensive plants run.
  • As of 2005, in Europe we have an emissions trading scheme, which is a market for CO2 credits.
  • By limiting the available credits, CO2 emissions become costly, and they are reduced.
  • In order to see the effects of a CO2 price, we added it to the model.
  • An investment decision is very much bound by uncertainty: it is based on expected power prices, expected CO2 prices, expected fuel prices, technology developments, company profile.
  • Now we end up with a model with existing power plants, the owners that operate them, the market to sell the electricity to, a CO2 market with limited credits to make sure there is a CO2 price that reduces the emissions, and the owners invest in new capacity as they see fit.
  • With the model, we now study the influence of policy on these investments in the electricity market, the CO2 emissions that result from this, and therefore also the demand for CO2 credits, the CO2 price that results from that, and so on.
  • The CO2 market works fine on the very long term, but it is not unlikely that it leads to a rather expensive decarbonization path.

Return to Summaries

(image source)

 

Leave a Reply

Your email address will not be published. Required fields are marked *