Assignments WS 2012/2013

A local restaurant during lunchtime

In the place I live is a restaurant which serves lunches during lunchtime. There are places for 40 visitors, 10 tables, 1 kitchen and 1 bar, 2 waiters, 2 chiefs, 6 items on the lunch menu and 10 items on a beverage menu.

Entities, attributes and constraints:

Visitor

• he has to decide what he wants to order (more items on the menu = more time to decide)
• can order any item from the menu and any item from the beverage menu
• he's got just 30 minutes for the lunch because he works and his time for the lunch is limited
• after 30 minutes he's getting angry
• prefers seat alone at the table (see Table bellow)
• if the restaurant is crowded (or full) he could be unsatisfied

Waiter

• receives orders from visitors
• put orders to the kitchen
• serve orders from kitchen to the table (visitor)

Cook

• works in the kitchen
• prepares lunches
• the more kinds of lunches, the more time to prepare or cook the lunch

Table

• provides 1-4 places for visitors
• they are placed in the restaurant and each of them has different distance from kitchen and bar
• more tables = more places for seat = more visitors

The simulation goal is to set a real model of the restaurant and try to find a better optimum of menu items/waiters count/cooks count related to customers satisfaction.

Tomas, what's now? :)

Thank you.

Jiří Hradil 10:23, 5 December 2012 (CET)

OK, what simulation tool would you prefer to use for it? Tomáš 11:21, 6 December 2012 (CET)

Sorry Tomas, I'm going to use Netlogo for it. May I understand it that you have accepted the simulation? Thank you. Jiří Hradil 11:33, 6 December 2012 (CET)

I think Simprocess would be much better platform for this. I don't see there enough true "agent" characteristics. If you are willing to make it in Simprocess, it is accepted. Please, try to keep as many real parameters as possible. If you still would prefer NetLogo, please, try to explain what benefit do you see in making it in this tool. Tomáš 14:11, 6 December 2012 (CET)

In Netlogo I can simulate avoiding of waiters if they cross themselves, avoiding guests, compute distances between waiters, tables and kitchen using vectors, simulate walking in space and so on. For example more guests will slow down the restaurant because they can run into others and so on. I think that Simprocess is not a right tool for it because it can't describe reality as it is. I really hope that you trust me that I will do it in the right and proper way. Jiří Hradil 14:26, 6 December 2012 (CET)

All right, but bear in mind that if you stipulate it this way, it means a simulation of agent movement, avoidance, etc. etc. It could be a pretty complex problem with several hidden pitfalls if you want to make it realistic. If it is OK for you, then accepted and good luck.Tomáš 14:58, 6 December 2012 (CET)

Hive

author

--Adam Sedláček (xseda07) 12:09, 30 November 2012 (CET)

As is known, the hive is a community that is very complex and is influenced by both internal and external factors. Therefore, I propose to create a simulation that would examine the influence of internal and external circumstances which influence the hive and the ecosystem in which they live. Especially I mean these factors:

• Number and distance of flowers and fruit trees to pollinate
• number of workers and drones in the hive
• how early beginning / end of summer / winter affect stocks of honey and colony survival
• how number of workers affect pollination of trees/ flowers and therefore the number of flowers and trees in neighbourhood

How it will work

Workers are responsible for getting food for whole hive and also for pollination of trees and flowers. They are using sun for the orientation and they communicate between themselves. In the beehive there ale drones, which fertilizes Bee´s queen, bee´s queen, which oviposits and the eggs-larvas, which are later transformed to either workers or drones.

Trees/flowers need to be pollinated in order to proliferate. There will be also change of season and different hive behaviour during summer/winter

Method I will use NetLogo to simulate the Hive

OK, sounds interesting. Accepted. Please, identify yourself. Tomáš 22:49, 29 November 2012 (CET)

I am sorry. --Adam Sedláček (xseda07) 12:09, 30 November 2012 (CET)

To vote or not to vote?

author

--Riccardo Torchio) 15:53, 3/12/2012 (CET)

I want to design a simulation that can show us how people decide if to go to vote or not.

I will create agent with -proximity to a person that can influence the agent -willingness to go to vote -expected result of the vote I can take my conclusion about the threshold value of this 3 variables in this process.

Method i will use NetLogo to simulate the process

Generally it could be a good idea, but you should elaborate it into a greater detail. Since it is a "soft" problem, it is ambiguous. How it will be connected with reality? How you can evaluate your results? What you are going to measure? If you stick to this topic, the best idea would be to find a scientific article or an experiment where similar problem is explored and to design your task accordingly - on its basis.

If you just create what you have suggested, well, you will have a nice NetLogo model. What's the point? Please, think twice about it.

Tomáš 12:16, 4 December 2012 (CET)

Default

author

--Riccardo Torchio) 18:34, 4/12/2012 (CET)

topic As a second proposal i'm thinking about a simulation of default by a state. My agents will be: - banks -citizens

The news of the dafault could be placed in a randomly around. At this point the agents that enter in contact with the news will run to bank to withdraw theyr money. If someone see queue at the banks will help the process to speed up, entering the queue. I can take some conclusion about the self-fulfilling prophecy of a default by banks and how people react to this news.

Method i will use NetLogo to simulate the process

Bank runs are generally a perfect phenomenon to simulate. I like the idea, however I don't see there a major benefit to make it as agent based simulation particularly in NetLogo. What I would recommend is to wait until you first systems dynamics class (tomorrow I guess) and to try to consider creating it using Vensim. Tomáš 15:08, 6 December 2012 (CET)

--Riccardo Torchio) 15:29 7/12/2012 (CET)

I'll not be present today at lesson but i will download the slide asap and surely i'll follow your advice, thank you. so it's approved?

I think this could be suitable simulation for the Vensim. In this case it is not just about spreading the bad news, but about all different factors. You just need to put this simulation into some real context. Take for example the Greece. There are people withdrawing their money or sending them abroad and in the opposite direction the EU is sending the money to Greek banks. And you can think of other factors. So if you place your simulation of default into the Greek context and work it out the direction I have suggested, I would accept it as a Vensim simulation. Oleg.Svatos 19:08, 8 December 2012 (CET)

--Riccardo Torchio) 15:25 9/12/2012 (CET)

ok so i will start to work on this, thank you

Enter the market or not?

I was thinking about a firms competition simulation. A first firm decides to enter the market or not and depending on the choice, the firm already in the market decides to fight or to accommodate the new firm. For example, I could take the example of fast-food in Belgium: the burger king chain could be trying to enter the market and thus the McDonald's chain already present could decide of its behavior.

Wagon Antoine Qantw00 22:10, 3 December 2012 (CET)

Good idea, Antoine, but you should elaborate it into a greater detail. It is too vague so far. What will be the entities, how you will treat products, how you will work with prices, what will be a role of customers...?

I really like the thought, however market simulations are always complicated, because people usually tend to add more and more details. You should imagine, how it should work and find a reasonable and well defined situation what you will simulate.

You can also wait for you systems dynamics class. Perhaps it could be a proper approach for your idea.

Tomáš 12:30, 4 December 2012 (CET)

Basically, here is how will work my simulation (can be improved in the future of course):

• At the beginning, there are only McDonald's firms on the market. They decide on the location, the price and the quality of the products.
• The customer decides to go to a restaurant regarding the location, price and quality
• At the second step, Burger King decides to open a restaurant or not. if yes, it has to choose of course the location, price and quality.
• They, when there are the two brands on the market (McDonald's and Burger King), McDonald's will decide if they fight or accommodate. if they find, they will either provide a lower price or a better quality for the same price. if they accommodate they will change their features in order to get the same as Burger King and thus split the market.
• Both brands can also decide to open new restaurants.

The goal for each brand will be to get the best market share.

I think that Netlogo is the most suitable platform for this simulation. I don't know if it is realistic and feasible; I just tried to gather some ideas for the simulation. Thus what do you think about it?

Wagon Antoine Qantw00 19:31, 6 December 2012 (CET)

In this case you should deal with two things - first, unlike our "market structure" example, the product is not homogeneous. People usually do care whether they eat in McD or BK and prefer one of them. Some people do not care at all, some people prefer one of the brands up to the extent they will never buy the second one. The majority will probably prefer one over the other, but will have no problem to eat in the other chain if it is more convenient in the moment. Second, you will have to deal with proximity. I.e. people simply prefer visiting closer restaurants over the farther ones. You should incorporate it into your model, but please avoid random "awarming". If it is fine for you, then it is accepted. Tomáš 12:46, 8 December 2012 (CET)

Disease

author --Pilar 10:57, 4 December 2012 (CET)

Subject of simulation

I would like to simulate spread of a disease which would take incubation period (i.e. person is acting as carrier of the disease but doesn't have any symptoms) and immunity into account as suggested in Modelování a siulace komplexních systémů by Radek Pelánek. The simulation will be based on SIRS model of diseases with constant population (simulating period of weeks or months and non-lethal diseases).

Objectives

To determine influence of relative length of incubation period, duration of disease and duration of immunity on the dynamics of the disease in given population - will any patterns develop, when do more or less people become ill, can the disease die out in given population?

Method

MAS with random contact of agents.

Entities and properties

Agents in the model are people. Each person has a state (healthy, infected, ill, resistant), information about progress of the disease in his case (how long he will stay infected, ill or resistant), whether he's been vaccinated or quarantined. Agents move around randomly. When agents come into contact and one of them is infected or ill the infection can transmit to healthy agent. Resistant agent cannot become ill. Global parameters include population density, mean length and standard deviation of incubation period, illness and resistance period (which are instantiated for each agent when he changes his state appropriately), general susceptibility, what portion of the population has been vaccinated, efficacy of vaccination and severity of quarantine. Vaccinated agents have lowered chance of becoming infected (the chance is given by the efficacy of vaccination) when in contact with infected or ill agents. If quarantine is applied ill agent's contact with other agents is limited. The model will simulate this effect by lowering the chance of transmitting the disease from quarantined agent to healthy agents.

First, I think it was already used in the past, and second: it is better to create your own original assignment. However, you definitely can use it as an inspiration and redefine it into something new, what would be derived from this idea.

Tomáš 21:23, 4 December 2012 (CET)

Alright, model would enable comparing effects of different actions on the dynamics of the disease (vaccination, quarantine with different strictness, e.g. complete quarantine versus home rest with contact with fewer people) as well lengths of phases of the disease. I'd like to see whether any stable patterns emerge in such environment and what are the limits of disease survival. I haven't found a similar model and it's not elaborated in the mentioned book, it's just suggested as an idea for extending the basic virus model.

Pilar 15:03, 8 December 2012 (CET)

OK, please describe model details: kinds of entities, their basic properties and interaction and how it should look like. Tomáš 16:03, 8 December 2012 (CET)

OK, described above

Blood type

Author --Marta Machová 10:12, 5 December 2012 (CET)

The most important blood-group system is ABO system. Its named after 4 blood group (each is result of composite of two allele):

• 0: 00 (=homozygote)
• A: AA or A0
• B: BB or B0
• AB: AB (=heterozygote)

Evolution of blood groups

Each of blood group (except 0) where evoluted from "0" type by adding antigenes A, B and combination of both. The system of inheritance blood group have exact rules by accepting one allele from each parent. Because of evolution, the groups are increasing and decreasing the amount of its representative. Id like to predict, based on multi-agent model simulation, the future statistics and possibility of disappearance one of group (0).

How it works

There is six possible combinations of two allele 00 (bloodtype 0), AA (bloodtype A), A0 (bloodtype A), BB (bloodtype B), B0 (bloodtype B) and AB (bloodtype AB). Each allele is given from one parent. So the inheritance of bloodgroup is very clear but on the other hand, for exemple, when mother have 0 type and father B, their child have two possible type: 0 or B, depends on what combination of allele made fathers type. It could be BB or B0, so child will have one of: 00xBB: (0B,0B,0B,0B), 00xB0: (0B,00,0B,00). The interesting thing on this exemple is, that in general we are expecting, that the result will be "B" type because "0" means: "have no antigen A or B" and B sounds like more dominant, more progressive. But in fact, when we dont know whether the B type of father is combination of 0B or BB, we could be surprised that 0+B=0. If we rewrite this into "pairs of alleles" it will be clearer: (00)+(BB)=100%(0B) and (00)+(B0)=50%(00)+50%(OB). It was exact example only to show how it works in detail-easy and clear.

Different distribution of ABO

There exist ABO blood type distribution statistics by nation, by ethnics, by continent etc. I like to simulate next development of amount of people with each bloodtype and detect, if there is a trend of dissapearance of "0" type.

Method

Id like to use NetLogo to simulate the evolution of blood groups --Marta Machová 10:19, 5 December 2012 (CET)

I think that for this kind of simulation (based on evolution) its suitable NetLogo tool. Agents of four blood type need to meet to give one allele that is base of two alleles gen. There will be no evolutionary pressure except the locality. The agents will not move all the time everywhere on view. They will stay with big probability on one continent. The View will be devided into continents with defined percentage of each blood type. It depends on detail I need to simulate but its also possible that individual countries are need to be defined (the statistics data exists). And individual ethnics could be defined also.--Marta Machová 03:44, 7 December 2012 (CET)

Agents and parameters

• Woman-gender(w,m) alleles[(0,A,B),(0,A,B)],bloodGroup (0,A,B,AB), age(0-60), colour(four-each bloodgroup one), children(0-3)
• Man-gender(w,m),alleles[(0,A,B),(0,A,B)],bloodGroup (0,A,B,AB), age(0-60), colour(four-each bloodgroup one)

Behavior

Agents are randomly moving, when meets two different genders and in case age of w is between 25 and 35, age of man is between 20-50 and woman have less than 3children, born child with random combination of alleles of parents.

The age siutable for having a child is changable as well as number of children (based on datas from Statistics office-average age of having child, average number of children), before start of simulation its also required to set the percentage of each bloodGroup.

---

I like the idea, however from my point of view systems dynamics with Vensim would be perhaps more convenient tool for this. Oleg, what do you think?Tomáš 11:24, 6 December 2012 (CET)

I think, that the multi-agent approach is more appropriate. As I understand it, this simulation would be simulated by many agents whose property would be the different blood type. So it is not about relations among several entities, but about greater number of entities which randomly meet and create a new entity with some blood type. Oleg.Svatos 18:24, 6 December 2012 (CET)

I do not see any need for meeting of anything there. OK, perhaps it is not clear. Marta, please, could you elaborate your assignment a little bit in order it would be clearer how do you plan to solve it? Tomáš 00:22, 7 December 2012 (CET)

I have just added some more description of what Id like to simulate, how exactly the system of inheriting works and some more model method details --Marta Machová 03:44, 7 December 2012 (CET)

OK, that't pretty interesting. And now, how you will simulate it? What will be the entities? What relationships will they have? How they will interact? What interaction they even need? Tomáš 13:18, 8 December 2012 (CET)

I have added description of agents and its behavior--Marta Machová 09:43, 9 December 2012 (CET)

Ok, accepted. Tomáš 13:16, 9 December 2012 (CET)

Landscaping the areas

Landscaping workers create new functional outdoor areas. Their duties include planting bushes, trees, sod, and other forms of vegetation. Requirements of the project: number of planted stocks of trees, bushes etc. (or m2 of planted area). The ideal time to plant trees is early spring before budbreak (number of days or months).

Elements of this model: experienced and new workers, productivity of workers, assimilation rate, planting rate, different overhead parameters (training new workers, weather conditions etc.)

The goal of the simulation: Simulating and testing the model, set the different parameters, comparison of results and finding the optimal solution. I'd like to use as an inspiration model "Brooks’s Law Model".

Method: Vensim --Achatov Igor 14:00, 5 December 2012 (CET)

I don't exactly get who would be this simulation for (who the target user is) and what are exactly the parameters it would help to set up. Oleg.Svatos 19:00, 6 December 2012 (CET)

This model could be useful for the project planning process (e.g. for project manager). It will help to complete the landscaping project in time (with a limited number of days), to regulate the planting rate and productivity with dependent constraints, such as: hiring and training of new workers, the influence of weather conditions (heavy rain, strong wind, and hail), elapsed time etc. If you want I can add additional parameters like material (for example planting stocks, fertilizers), financial ratios (e.g. salaries, material costs, profit - based on real costs and prices). The regulation of these parameters can be useful for forecasting financial and economic activities, cost control and productivity. Achatov Igor 10:08, 7 December 2012 (CET)

OK. Include the additional parameters and it is accepted. Oleg.Svatos 10:12, 8 December 2012 (CET)

Fission-fusion society

This form of social organization occurs in several species of primates (chimpanzees, bonobos, ...). These societies change frequently in their size and composition due to changes in their environment and/or due to individual animal dynamics.

Elements of this model: primates, sources of food, predators

The goal of the simulation: Simulating societies and testing how they react to changing conditions (parameters). Also compare model to real life data (behavior of animals/groups, size of groups).

Method NetLogo Xpalj 24 20:15, 5 December 2012 (CET)

OK, it could be... Please, just describe in more detail how the simulation should look and work. Tomáš 15:42, 6 December 2012 (CET)

Market in MMORPG EVE Online

Author --Xbecj07 20:40, 6 December 2012 (CET)

                Since making simulation of real market is too ambitious and results would be useless due to inevitable over-simplifying of problem, i would like to simulate market which is by itself simplified.

EVE online is online game which is based on player vs player interaction. This interaction happens also in form of Market transactions, so this brings many aspect which are similar to real life markets, but still it is much less complex. For example bubbles, different prices for same commodity on different places (markets) etc.

Entities and attributes:

Market

All transactions could be held only on market. Agent places sell orders on market for price,  he chooses.  Market stores these orders and in case of transactions discardes paired orders. Consumption happens on markets. Production of commodities happens on some markets.

Sell order

Selling agent places order on current market. This order is always set for price, which is smaller than current smallest price. Each order is paired with agent and in moment of transaction this agent recieves payment. Simulation could possibly contain time of order placement and discount old orders, by rate defined by agent’s variable.

Commodity

Tradeable good. Each of good have some maximal amount which could be transported by agent in one moment. In simulation, there could be only one or possibly more commodities.

Consumption

Consumption represents in-game consumption of products by players, which does interfere with Market, but whose in-game actions are not focused on trading. To make this aspect more real, consumption should bigger on markets with higher capitalisation (players are more likely to buy on bigger markets).

Agent

There are two basic types of agents.

1. Non moving agents – consumers and producers. These agents will be (probably) represented by enviroment (actions on market). Producing agents will place sell orders on their market. They will put back their cash (from sold commodity) on market (in form of consumption).
2. Moving agents – traders

These agents move commodity from one market (with low sell price) to another (with high sell price). Each of these agents will have his own „Wallet“ with amount of cash and set of his own market orders. Each agent must pay certain amount of cash, otherwise he will dissapear, together with his orders. Each agent will trade only one type of commodity.

 For running simulation, Netlogo seems to me as a best tool.

OK, what it is good for? If you want to simulate market, it is a better idea to avoid the simulation of the whole market, i.e. the whole economy and focus just for a certain aspect. In our "market structure" example we have focused on a market with homogeneous product just after release of monopoly, It is enough specific situation that it allows a reasonable simplification without becoming useless (and the results are more or less plausible then). I would recommend you something similar. Try to find certain very specific market situation and try to create a model accordingly. It should be much more useful than to create a model of a model. Tomáš 13:30, 8 December 2012 (CET)

Well i wanted to watch how products flow over the area and how fast do price change affect markets with smaller volume (how long does it take for markets in area to adjust to sudden change of price (sudden flow of products) to area. Xbecj07 16:27, 8 December 2012 (CET)

Using and starting market sentiment for profit

Author --Xbecj07 16:28, 8 December 2012 (CET) Well, this is my second idea. There is one market with more products on which we have standard buy and sell orders. Agent can change these orders. Point of simulation is to find how can one agent benefit from sentiment on market (so called bubble). There are more players on market, they got different possibility to be affected by short term changes on market (sentiment). Also these agents choose which product should they trade (from two possible). Point of simuation: A) Determine necessary share of market volume of agent to enable him start sentiment and profit from it (for example he sells huge amount of commodity, thus lowering the price and starting short (sell) sentiment, which enables him to buy more of commodity later for much smaller price. B) Determine how does sentiment on one market affect sentiment on other market (some agents are active on both market and their amount of money is limited. Thus these markets should affect each other. C) Determine which combination of aspects is most profitable.

 For running simulation, Netlogo seems to me as a best tool.

OK, market bubbles are a good material, but as you have stated, it is very general. I would choose one particular product and preferably one particular historical example. This way you can use hard data to evaluate its fit with reality. For this kind of problems, both agent-based simulation and systems dynamics are possible, it depends heavily on the particular on the particular specification. Tomáš 14:00, 9 December 2012 (CET)

IT seems definitely better to choose only one commodity. I think silver could be good one, since there was recently small bubble on this commodity. Basically i can see two types of simulations, which would be run and results would be compared.

A) Market where players are marginal (testing agents against real life data, which is not changed by our agents). Real life data will be taken from Patria.

B) Markets where players are mayor (market is result from actions of agents). Xbecj07 17:49, 9 December 2012 (CET)

Virus containment

I have seen many disease spread simulations but no solution to this spread. So I the simulation of steps taken by organisations like WHO in cases of a virus.

Entities:

• Healthy people
  • every person has soome basic attributes such as: immunity, health points, carrier_potential
  • people can become carriers of a virus or infected
• Sick people
  • sick people loose their health points based on virus properties and the general spread (this effects also the efficiency of the treatment)
• Doctros
  • they try to heal people
  • they themselves can become carries or infected
  • infected doctor's work performance decreases
  • After succesful treatment of most of the patients, surplus doctors can leave the area (this can cause another outbreak when not managed carefully)
• Virus
  • viruses are defined by their spread factor and deadliness

Focus of this simulation:

• Number of the doctors required to contain different stages of virus outbreak.
• Number of doctors that are not needed any more (the people management and deployment problem - when to deploy more, when to send doctors home due to surplus)
• Simulation of diferent reaction times to the virus outbreak.

Implementation tools:

• NetLogo

Jakub 10:28, 7 December 2012 (CET) (xstaj68)

Jakub, if I understand it well, you want to simulate a local epidemy of an infectious disease like e.g. ebola or the like. You have one entity with two possible states: either healthy or sick and its specialization, i.e. doctor. Or do you consider virus an entity as well? So, first you really should choose a specific illness in order to work with real parameters (you will need to perform a certain research on it). Second, how do you suppose the agents will interact, in other words, how the simulation should look like? Tomáš 15:56, 8 December 2012 (CET)

Tomáš, I was thinking about something like this: entity Person(states: healthy, carrier=spreads the virus, sick=spreads the virus and is being killed) and specialisation Doctor(same states -> influences ability to help others). Virus is an abstraction, that can be modeled into some real disease (by setting its parameters like how fast it spreads=influenced also by immunity of person and how deadly it is=how many steps it takes to kill a healthy person) - I dont think we need to tailor the virus down to particular case. Simulation would start with setup - randomized/configured population, spread of virus, local doctors. What will drive the simulation is an ability of WHO to react quickly (delay between deployment command and actual arrival of the doctors) and deploy the proper amount of doctors (on click event). We will also be able to see if deployed number of doctors were successful in their affords and we can setup also the minimum required number of doctors to be deployed = in case of successful treatment of the big part of the population some of the doctros can be send home(can be parameter) - what can cause a second outbreak. Jakub 16:23, 8 December 2012 (CET)

First, I would really prefer if your model will simulate a particular event, particular epidermic. Of course, it should be designed generally, however it should work on very specific parameters. We almost always require this way, because it means the need to invest significant amount of time into an underpinning research, the paper is much more valuable then and it is the only way, how to verify the model. Second, and it is the biggest flaw I see here: the role of medical personnel in case of WHO operations is usually just to help stop spreading the disease by hygienic arrangements and to provide supportive treatment, but they typically really do not heal the sick as we aren't able to treat the majority of the infectious diseases. You should incorporate this fact into your simulation - and by the way, this is one of the reasons why it is desirable to solve a real situation, since you will be able to avoid such flaws. Tomáš 14:17, 9 December 2012 (CET)

Tomáš ok I thought abou it and i figured more interesting simulation, please see: Schoolig of the fish Jakub 15:10, 9 December 2012 (CET)

Two Ice Cream Vendors

Consumers are randomly distributed along a let's say one mile long beach. They all like ice cream the same and dislike walking the same. Prices are regulated and equal for every vendor. There are only two vendors and each is on the other side of the beach. That means that both have almost same profit - half customers go to left side and half to right side. They decide to take customers from the other vendor by moving to another position.

Entities/units model:

• customers - randomly distributed, moving ramdomly, only preference is distance
• vendors - same prices, can change position by user, then simulation starts from begining again, count visits (=actual purchase of ice cream)

Simulation purpose:

• Model will simulate changes in profits of vendors by removing stand to another position. And try to find the optimal position with maximum profit.

Tool

• NetLogo simulation model

--Overfloater 13:38, 8 December 2012 (CET) (xtvrp03)

Good, please, include also the following situations as options: more than two vendors, vendors cooperating/noncooperating, non-fixed prices. Accepted. Tomáš 14:08, 8 December 2012 (CET)

Schoolig of the fish

Aim of this model is to emulate schooling behavior of fish. There will be only one entity: fish. Simulation starts with randomly distributed fish and these fish behavior is constraint by the same set of rules from which schooling will emerge.

Fish:

• has some movements attributes: turning angle, speed of movement
• has tendency to join the school
• has tendency to break apart from school
• has awarness of other fish, so it can prevent colisions

Jakub 15:09, 9 December 2012 (CET)

OK, good, and the goal of the model? What should be your research report about? Tomáš 15:22, 9 December 2012 (CET)

We could focus on a leadership. When a school emerges, to the observer it seems that one particular fish is leading the school. We try to predict emergence of leader by studing the initial positions of fish and measure stability of the leadership. Jakub 15:54, 9 December 2012 (CET)

We can also measure the leadership potential of each fish in given step Jakub 15:56, 9 December 2012 (CET)

OK, please, explore the relevant existing literature and discuss in your report how your model fits reality. Accepted. Tomáš 15:59, 9 December 2012 (CET)