Difference between revisions of "Assignments WS 2013/2014"
(→Political compass) |
m (→Fertility and intelligence) |
||
Line 40: | Line 40: | ||
::The agents will be moving around on the grid representing the world, their movements shall be random. Each time agent meets another agent, several attributes will be evaluated e.g. age, marital status, maybe some other attributes → these attributes decide if the agent '''can''' find life partner (e.g. age < 15 - too soon, already married - bummer - I don't want to introduce divorces - it would be too complicated then). Then, based on iq/fert value of each agent (again, I will be abstracting from common hobbies etc due to the complexity of the model), a probability that the two agents become life partners is evaluated (people with similar IQ have higher probability that they will "understand" each other), as a last step, the probability is used to decide whether they are going to be the life partners or not... It may happen then that a genius won't find partner at all in a population of dumb people. If you were asking me for the particular pattern to be used to derive the probabilities, that I don't have yet. For that, I need to evaluate or estimate some key parameters (how many people can one meet in a whole life, how many of them is suitable for being a partner, etc...). Vojtěch Zrůst, [[User:Xzruv00|Xzruv00]] 17:56, 10 December 2013 (CET) | ::The agents will be moving around on the grid representing the world, their movements shall be random. Each time agent meets another agent, several attributes will be evaluated e.g. age, marital status, maybe some other attributes → these attributes decide if the agent '''can''' find life partner (e.g. age < 15 - too soon, already married - bummer - I don't want to introduce divorces - it would be too complicated then). Then, based on iq/fert value of each agent (again, I will be abstracting from common hobbies etc due to the complexity of the model), a probability that the two agents become life partners is evaluated (people with similar IQ have higher probability that they will "understand" each other), as a last step, the probability is used to decide whether they are going to be the life partners or not... It may happen then that a genius won't find partner at all in a population of dumb people. If you were asking me for the particular pattern to be used to derive the probabilities, that I don't have yet. For that, I need to evaluate or estimate some key parameters (how many people can one meet in a whole life, how many of them is suitable for being a partner, etc...). Vojtěch Zrůst, [[User:Xzruv00|Xzruv00]] 17:56, 10 December 2013 (CET) | ||
− | :::I am not very happy about this "random movement" way of simulation in NetLogo. It is quite common, because easy, however often pretty misleading. In fact, who you meet is not completely random, because you tend meeting people who are similar to you. Similar hobbies, similar interests... Of course, there is always some randomness, perhaps a lot of, but your approach isn't very good in this point. Please, try to reconsider it. Otherwise I don't see any problem. This could be a very nice assignment. | + | :::I am not very happy about this "random movement" way of simulation in NetLogo. It is quite common, because easy, however often pretty misleading. In fact, who you meet is not completely random, because you tend meeting people who are similar to you. Similar hobbies, similar interests... Of course, there is always some randomness, perhaps a lot of, but your approach isn't very good in this point. Please, try to reconsider it. Otherwise I don't see any problem. This could be a very nice assignment. [[User:Tomáš|Tomáš]] 22:59, 11 December 2013 (CET) |
==Fire evacuation simulation== | ==Fire evacuation simulation== |
Revision as of 22:59, 11 December 2013
Please, put here your assignments. Do not forget to sign them. You can use ~~~~ (four tildas) for an automatic signature. Use Show preview in order to check the result before your final sumbition. |
Please, strive to formulate your assignment carefully. We expect an adequate effort to formulate the assignment as it is your semestral paper. Do not forget that your main goal is a research paper. It means your simulation model must generate the results that are specific, measurable and verifiable. Think twice how you will develop your model, which entities you will use, draw a model diagram, consider what you will measure. No sooner than when you have a good idea about the model, submit your assignment. And of course, read How to deal with the simulation assignment. |
In order to avoid possible confusion, please, check if you have added approved in bold somewhere in our comment under your submission. If there is no approved, it means the assignment was not approved yet. |
Contents
- 1 Fertility and intelligence
- 2 Fire evacuation simulation
- 3 Simulating and optimizing a process of opening an account at the branch of a bank
- 4 Political compass
- 5 From field to forest
- 6 HIV Infection spread
- 7 Intersection (vs. Roundabout)
- 8 Blue Eyes
- 9 Colonization
- 10 Optimization of public transport (entrance and exit)
- 11 Predační kaskáda - jak vyčistit přírodní koupaliště
- 12 General Behavior of Infections
- 13 Heat recuperation
Fertility and intelligence
Demographic studies have indicated that in humans, fertility rate and intelligence tend to be inversely correlated - the more intelligent (measured by IQ tests) a person is, the lower total fertility rate it exhibit. Survival rates are also correlated with IQ. Does this mean, that the humans get dumber than they are now?
To relax the problem, lets assume that there is a designed, shared space (capacity) for fertility and IQ. What would be the final ratio between intelligence and fertility?
Each agent in the model would have a complementary values for inteligence and for having kids (i.e. iq 0.7, fert 0.3). The agents first search for a life partner (there is a higher probability for the ones with similar iq/fert to became life partners - the fact is related to IQ). After finding one, the probabilities of fertility of both parents will determine if they will have a kid, or kids. If so, the kids will acquire parents' averaged values with some possibility of mutation. What will happen if we populate a hypothetic world with two types of people - the clever ones and the fertile ones. Will there be a minority of the clever people, or will they became extinct? Will there be only one group left with a certain amount of iq (higher IQ than within the fertile group at the beginning but dumber than within the ones from the clever group)?
I will use Netlogo. Furthermore, depending on complexity of the problem, I would like to introduce basic features like agent age and general life time events, concept of family, etc...
Vojtěch Zrůst, Xzruv00
- Generally, I like this assignment. Although it deals with a "soft" problem, what is usually tricky, you are approaching it right way, working with quantifiable facts, thus it is no problem. The only issue I see is how you will actually simulate finding of life partners? Please, explain. Tomáš 11:18, 10 December 2013 (CET)
- The agents will be moving around on the grid representing the world, their movements shall be random. Each time agent meets another agent, several attributes will be evaluated e.g. age, marital status, maybe some other attributes → these attributes decide if the agent can find life partner (e.g. age < 15 - too soon, already married - bummer - I don't want to introduce divorces - it would be too complicated then). Then, based on iq/fert value of each agent (again, I will be abstracting from common hobbies etc due to the complexity of the model), a probability that the two agents become life partners is evaluated (people with similar IQ have higher probability that they will "understand" each other), as a last step, the probability is used to decide whether they are going to be the life partners or not... It may happen then that a genius won't find partner at all in a population of dumb people. If you were asking me for the particular pattern to be used to derive the probabilities, that I don't have yet. For that, I need to evaluate or estimate some key parameters (how many people can one meet in a whole life, how many of them is suitable for being a partner, etc...). Vojtěch Zrůst, Xzruv00 17:56, 10 December 2013 (CET)
- I am not very happy about this "random movement" way of simulation in NetLogo. It is quite common, because easy, however often pretty misleading. In fact, who you meet is not completely random, because you tend meeting people who are similar to you. Similar hobbies, similar interests... Of course, there is always some randomness, perhaps a lot of, but your approach isn't very good in this point. Please, try to reconsider it. Otherwise I don't see any problem. This could be a very nice assignment. Tomáš 22:59, 11 December 2013 (CET)
Fire evacuation simulation
My intention is to create a model of evacuation process of real public building in case of fire emergency. My model considers people as agents who don’t know the way out and in case of fire, they have to find the way first. For this reason there are green tables on the walls of public buildings that should help people find their way to safety.
Agents will have to deal with fire and toxic smoke spreading through building on their way out and also other agents, who could crush them in panic.
For simulation I am going to use as environment 2D version of NetLogo 5.0.4.
My model will count the time of evacuation process and loses of people and help consider whether the green “Exit” tables are on right places and in right amount.
Tomáš Votruba, Xvott00 17:11, 7 December 2013 (CET)
- Generally, it could be an interesting research. It should incorporate more complex shapes of premises than just rectangle (in our class example). The greatest problem could be the intelligence of simulated people. If it should be realistic, it could be quite tricky to develop. Please, could you describe a little bit more in detail, how people would navigate along the green warning tables? Tomáš 01:58, 8 December 2013 (CET)
Every agent would have own line of sight dependant on where he is facing and how far can he see. When fire alarm will sound, agents will look around and try to find green table. If he sees one, he goes to it. Every green table shows direction (not strictly in line) to the exit, and agent will go that direction, until he sees another table, which will give him another direction, or exit. Green tables will work as some sort of checkpoints, which will lead the agents to safety.
If agent doesn´t see any green table, he will try to look for one by looking around and walking randomly chosen direction. My intention is to use one floor of real existing public building (for example) Centrum Černý Most or similar. I am not telling which one so far, because I still don´t know whether I get proper plans, I could use. But it will definitely won´t be just a rectangle.
Tomáš Votruba, Xvott00 15:37, 8 December 2013 (CET)
- OK, approved Tomáš 11:20, 10 December 2013 (CET)
Simulating and optimizing a process of opening an account at the branch of a bank
Currently I am working as a junior process manager at one of the banks in Prague and participating in a Lean Six Sigma project of optimization of an account opening by client visiting a bank`s branch. The goal of that project is to minimize time spent by client at a branch and the number of client`s signatures on various contracts by creating a unified multicontract form.
-
The goal of my simulation, which I`d like to make in SimProcess, is to simulate as-is process and identify, at which steps there is idle time and at which steps there could be a queue of customers. Afterwards I would simulate an improved process - that simulation should prove that the process is improved. The goal of simulation is the same as the goal of the real project that I take part in - it is to cut non-value added steps and reduce time spent by customer at the branch. Simulation will be based on real data.
xziny00 20:16, 1 December 2013 (CET)
OK, approved Tomáš 01:37, 8 December 2013 (CET)
Political compass
I am interested to create a simulation of the political spectrum in the NetLogo. The model would simulate the distribution of political views of hypothetical country and the development of leader´s preferences or political parties that people vote. There would be several variable parameters of the population, population growth, a possible radicalization of society, etc. The model would shows electoral preference at the time on the graph and there would be also moving agents on the graphics part of the model. One tick would correspond to one week. The whole model would therefore ran for four years, which is the standard electoral term. In this time there could occur some populist events that affect the electoral battle (floods, the introduction of direct democracy, etc.)
Political compass would be based on the known distribution of the political spectrum into 4 quadrants. The horizontal axis would represent the position of voters and political parties to the "left" and "right". The position on the vertical axis would show whether the voter or political party rather libertarian, or authoritarian opinions. Agents (voters) will change slightly their electoral preference every tick. This would simulate a real reflection of the political situation as caving scandals, bad / good economic performance, etc.
Tomas Lizner, xlizt00, 00:11, 3 December 2013 (CET)
- If this kind of simulation should be developed, I would think about systems dynamics instead of agent-based simulation. But, first and foremost I am generally not very happy about "soft" topics. It is very hard to measure such things like political attitudes. Often, the risk of similar simulation is that it will be nothing more than an academic exercise. Please, try to reconsider it. Tomáš 01:47, 8 December 2013 (CET)
OK, I have another idea of the model simulation in NetLogo. It would be a simulation of water flow throuhg Vltava water cascade. The principle of the simulation would be how fast can the individual dams release water without the river was on third flood stage (third flood stage is defined by flowrate). There would be agents: individual dams - they would have attributes such as inflow, outflow, retention, current capacity, the risk of rupturing etc. Between the dams would be used Links that represent the waterway between the dams (attributes: normal flow, current flow, flooding flowrate). These information I am able to find out through the internet.
There would be randomly generated a rainfall (or constant rain) at the beginning of the simulation and they would have an impact on the water cascade. One tick would correspond to one hour, during which time the water is slowly seeping into the ground itself. User would change sliders to make affect to the rate of outflow for individual dams and also there would be settings for weather. Changes in the running simulation would appear on the graph and in a simplified graphical display. I would use NetLogo, because simulation conditions such as a weather conditions and flowout rates of the individual dams can be changed in real time during the simulation, which makes the simulation more interesting and more relevant to reality.
Tomas Lizner, xlizt00, 14:11, 9 December 2013 (CET)
- First of all, I am sorry for not responding to your email. Before I managed to do it, you have offered a new topic. OK, I like this, the only thing to consider is if the best tool is an agent-based model. From my point of view, it could be done as a systems dynamics model in Vensim quite well. Netlogo simulation could be fine, however just in case, you will strongly rely on some spatial factors like maps of rainfall, etc. (and it could be quite demanding.) If this is not the case, I would focus on Vensim.
- So, please, specify if and why you really want to do it as an agent-based model or if you better switch to a systems dynamics model. As soon as we make this clear, I do not see any problem to approve this topic. Tomáš 11:34, 10 December 2013 (CET)
I would use NetLogo, because i like that idea, when user can change conditions in real time during the running simulation. For myself it is much more fun when I can see how parameter changes are instantly reflected on the behavior of the model. Another great advantage is that NetLogo better works with a graphical presentation of the results, which I consider much better than the results presented in some numbers. I do not plan make big complexity in maps of rainfall, becouse it very exceeds my knowledge, but there would be the possibility for expanding and improving my model in the future by another student. Another reason is the fact that , I have already made some model in NetLogo, so its environment is known to me and therefore it allows me to make better simulation of this problem with comparison of Ventsim. I also wanted to try create situation when after dams starts increasing outflow rapidly, threre will be tidal wave which for a while will increase the level of the river, which would be easier to create it in NetLogo.
- I am sorry, but the results you have mentioned are totally irrelevant. 1) Changing parameters of the model "on the fly" could be interesting for debugging or for playing, but it is totally meaningless if you should provide any relevant result of research. 2) Graphical presentation - the same. What is really important is numbers. You can present it any possible way using any possible tool (e.g. Excel), but you need to get to them using the most proper tool. 3) Your knowledge of NetLogo is fine, but it is no reason why to use a tool that is not the most appropriate for the actual problem. If you know NetLogo and don't know Vensim that much then hello! we are here to learn something new! 4) I am quite sure that tidal wave could be simulated in Vensim even easier.
- Really, I don't see there many reasons why to develop this simulation in NetLogo. I hope Oleg will agree. Please, either amend the assignment in any way that would make the use of NetLogo meaningful (maps of rainfall is definitely one option, but there could be perhaps some others) or develop it in Vensim. Tomáš 22:58, 11 December 2013 (CET)
From field to forest
When a field is left alone it takes quite a time until forest spreads on it; this process is called secondary succession. At first pioneering plants such as grasses and herbaceous plants grow there, later invasive small bushes and trees are following. Later on fast growing evergreen trees appear, but they are overtopped with deciduous trees and the forest starts to appear. I would like to propose a model of this situation focused on the influence of the fertility of the field in the very beginning of the process, i. e. I propose the question how long it takes for a forest to appear based on primary fertility of a field.
The process of succession is described well on Wikipedia: [1].
The simulation will be realized as a NetLogo model. Agents: herbaceous plants, invasive bushes, evergreen trees, deciduous trees. Agents parameters: probability of survival to the next year, probability to spread, shade provided, sunlight required.
Each agent can spread under different conditions, such as plants around (of the same kind and of other kinds as well), fertility of the ground and amount of sunlight avialable. Each plant lives for a specific period of time.
The fertility of the field is dependent on plants previously living there and on an initial value set in the model.
Alice Peková, xpeka00, 7 December 2013
- First, is the fertility of soil the only driver of the process? And second, I don't like that you would work generally with trees, bushes, etc. instead of particular kinds. I guess there are probably substantial differences in parameters (like the speed of growth) between different kinds of trees, aren't they? Tomáš 02:53, 8 December 2013 (CET)
The process is also crucially driven by the shade produced by plants. While the herbaceous plants provide quite insignificant shade, the deciduous trees produce so much shade to prevent other plants (like the bushes and evergreen trees requiring a lot of sunlight) from growing nearby. So I suggested some new parameters as you can see above (shade provided, sunlight required). However, I am not sure that differentiating between various kinds of plants would significantly improve the model, when the more general classes describe the behavior of the plants pretty well. I think that the variety inside of these classes can be simulated sufficiently using randomized values of their parameters. xpeka00, 11 December 2013
HIV Infection spread
Lucia Banakova, xbanl00, 9.12.2013
- I am more or less OK with this, but please, specify how you will actually simulate spreading of the infection. How will your simulation look like? Tomáš 11:38, 10 December 2013 (CET)
I would like to make the simulation as realistic as posible. A friend of mine is secondary school biology teacher and he promissed me help with the metodology. We will discuss posible transmission methods and their probabailities as well as factors that influence the probabilities. For example the main transmission route is sexual contact. When two agents meet, there is a probability under which they have sexual contact. Several factors contribute to this probability, including their marital status, sexual orientation and so on. If they have sexual contact and one of the agents is infected, there is a certain risk that the other agent will get infected. These risks are statistically quantified and I will use these statistics in the simulation.
Another methods of transmission are via blood or blood products and mother-to-child transmission during pregnancy, delivery or breastfeeding. The simulation will also take into account mortality (both HIV-related and non-related) and natality in the population.
The main parameters of the simulation will be size of the population, promiscuity rate and so on.
Lucia Banakova, xbanl00, 11.12.2013
Intersection (vs. Roundabout)
During the last year almost every day I go around an intersection at Střelnická / Třebenická which has a really unique shape which reminds me of a Roundabout. I'd like to try to simulate the Intersection as an agent based simulation. There are 4 exits from the road:
- to Kaufland (biggest local shop, about 130 parking places) - to Ďáblický hřbitov and local housing estate - to Kobylisy - to Střížkov
On every exit there are traffic lights with different timings, which I'll easily measure during one measurement at one specific time and day of the week. Most of the cars go straight on the main road from Kobylisy to Střížkov, they also arrive mostly at higher speeds than others leaving from the housing estate and Kaufland. I'd like to propose a simulation based model from NetLogo library with an intersection with 4 exists as is the real example. It will allow users to set frequency, deceleration, acceleration of cars from each direction. The lights will be set to constant ticks which will try to simulate the real scenario. I will omit the fact that cars entering the Kaufland parking place have to leave it also, it will behave as a normal street.
Then I'd like to try out a simplified model of Roundabout with the same parameters to compare which of these are better. The simulation would consider some safe distance before another car can enter the roundabout and will have same settings and parameters as intersection. Theoretically I'd love to have two models in one NetLogo project, but I'm yet unsure whether its possible.
Xklit17 09:58, 10 December 2013 (CET)
- approved Tomáš 10:27, 10 December 2013 (CET)
Blue Eyes
Every year we can hear some “scientists” talking about the extinction of blue eyes and blond hair, based on the assumption that the above traits are carried by recessive genes.
With the help of an agent based simulation, I want to study the inheritance process of blue eyes based on different scenarios where people will be segregated in different “countries” and each year (tick) some people will move to another country. Every country will have genetically different group of people, and the inhabitants of a country will only “merry” from their own country. I will try to monitor how long it takes in each different scenario for the blue eyes genetical trait to be significantly reduced. In each scenario the genetical composition of the countries will change, where the two extremes will be one where only one country has blue eyed people, and one where every country has the same percentage from every kind of eye colors (gene alleles).
The genetical inheritance model will be the simplified 4x4 Punette square matrix, not accounting for the fact that eye (and hair) color are changing throughout the life of a person. This is a simplified model which does not explain every natural occurrence, but can be easily used during the simulation. I will build my simulation in Netlogo.
If I find appropriate data I may even try to build a scenario similar to Europe, focusing on the Nordic countries where this genetical trait is the most apparent.
qhaud00 22:05, 10 December 2013 (CET)
Colonization
A simulation of colonization of previously uninhabited territory.
Simulation would run in netlogo. A new random territory is created at setup. Certain resource(s) and obstacles are placed on created territory. Settlers create settlements, use resources and try to create new settlements. Multiple different groups of setllers could be deployed and "fight" for territory and/or resources.
I would like to see results of this process depending on starting scenarios/conditions.
xpalj24 22:06, 10 December 2013 (CET)
Optimization of public transport (entrance and exit)
There are many public transports, where doors are supposed for entrance and exits at once. Unfortunately some people dont give priority to exit before entrance. So there can be a lot of delays. In some cities there are marked doors for coming-in or coming-out (for example London). There is another example Litvínov, where the doors closed to driver are supposed for coming-in with ticket check and rests are for coming-out, because of reducing numbers of black passengers. Is some solution more effective than others? It is question for my optimization and my simulation
Simulation will be implemented in netlogo (space will be divided into inside and outside of public transport with agents as people). User can set number of people, number of doors, throughput of doors (coming-in, coming-out or both). Simulation target will be decide, which option is the most effective.
I wanted to create this simulation last semester, you were ok with that, but I havent finished it until the end of last semester, so I hope It wont be a problem. Filip.hrbek 23:50, 10 December 2013 (CET)
Predační kaskáda - jak vyčistit přírodní koupaliště
Zaujala mě poptávka na vytvoření aplikace, která přišla k nám do firmy a rád bych se tímto zabýval i v rámci simulačního projektu, pokud je to možné. Níže (viz odkaz) je dobře popsáno o co jde. Aplikace resp. simulace doposud vytvořena nebyla, jednalo by se o prvotinu.
Díky, Ondřej Cahlík
Xcaho01 17:26, 11 December 2013 (CET)
General Behavior of Infections
My idea for semestral paper is visualization of spreading diseases amongst general population - in city or rural areas and also between cities or major population centers. Although humans behavior is „soft problem“, in this case it is quite deterministic, since virtually all of us have no desires to get infected by harmful viruses or to die. Especially if death is involved, virtually everyone is willing to go some length to protect themselves. I would like to use Netlogo, as it is most suitable mean for modeling this case, that is at my disposal.
At the start we will setup number of people, number of population centres (towns) and number of infected people. There will also be variables for setting, how will be the disease. After that, disease start to spread according to variables, which we set. If disease spread by physical contact, agents have to be on the same patch, if by air, there need to be one or less patch between them. All methods of spreading are quite obvious and standard. If disease reaches critical level, government can issue an evacuation order or shut down all traffic to prevent spreading the disease. Traffic between centres can be by trains or cars. If population centre is big enough, it has it´s own public transport, which will help disease to spread.
Setup variables:
- Number of people. - Number of population centres. - Number of infected people at the start.
Other variables:
- Incubation period length. - Chance to become infected. - Method of transfer (air, physical contact and possibly other method, like blood). - Average life remaining after being infected. - Time until virus become contagious after being infected. - Time until the virus manifest itself to our surrounding. - Chance to be cured. - If people can catch the disease after being cured.
I am not sure about difficulty of modeling, but I would also like to see in my model modeling of long term diseases in span of generations. There would be a fertility rate variable, chance of spreading disease form mother to unborn children, death by age and also by accidents. So there will be another variable, which will tell us, if model is for long term model or for shorter.
--Xvism00 18:49, 11 December 2013 (CET)
Heat recuperation
My idea for simulation comes from my brother's bachelor work. In this work he is trying to analyse possibility of recuperation of energy from waste water (basically said there is some energy in the water stored as heat which might be used to create electricity).
For this purpose my brother and his schoolmates collected some data about water used in the house. Unfortunately they were unable to collect data about temperature of the discharging tube which is a crucial piece of information for the project.
From collected data we can see how is changing water flows during day, week and year so we can approximate usage of the water by adding statistical methods. I would like to make a model which will show discharging tube temperature during a year.
Setup variables:
- mean temperature decrease while the water is used - approximation of water flow during day, week and year (trend of real data) - temperature decrease while recuperation device is on
It is possible that model variables will change during its creation but I suppose that adding new variable is much more probable then removing one.
I believe that Excel Monte Carlo simulation will be a good choice (theoretically there will be 52 columns (weeks) and maximum of 1440 (assuming thet every line is one minute) per each calculation.
xmalf00 Filip Malý Spiglnigl 21:16, 11 December 2013 (CET)