Hi,
I came accross somewhat interesting article on simulating crowd behaviour in conflict situations:
http://jasss.soc.surrey.ac.uk/4/3/7.html
I'm not convinced that you can simulate these kind of events and I don't understand what is the purpose of this simulation. Authors point out quite obvious fact that the more "hardcore", as they call them, participants, the more probability of rising aggression and a fight.
I realize that you cannot invoke riots to analyse crowd behaviour in stress situation, but what do you think about simulating crowd behaviour?
What do you think about the behavioural model of the agents proposed in the article? Have you ever participated in a bar fight, football match brawl or maybe even riots?
To be honest, I barely see the purpose of this kind of software. I think science should concentrate on software to predict and avoid conflicts. Do you think that some kind of human motion analysis software is able to detect possible conflicts or "hardcore" participants in the crowd? Do you know any software of that kind?
It seems to me that the research of the crowd behaviour understood in general term each year is becoming more and more popular. Behavioral models presented in the article should not surprise anyone. It was predictable that sooner or later someone would come up with the idea and write such article. Subject seems to be quite interesting in the context of the current geopolitical situations in Europe and the world. However, I have serious doubts whether such simulations have any take-up in the real world. This type of simulation requires a vast number of different factors, and one does not have to be a genius, to know that with large enough number of variables any model can be a) unreadable b) different from the reality, hence the reasonableness of such simulations maybe reduced to zero. Unless someone just want to write yet another article to add to his scientific achievements :). Simulating crowd in non-conflict situations is a completely different matter. The crowd in suchsituation is more predictable, hence the conclusions drawn from these simulations are more relevant. Everybody behaves differently in conflict situations, and even if we claim that we would never have done something, once in a conflict situation it may turn out to be an empty promise or misplaced hope. I never participated in such events as bar fights, brawls at football match or riots, so my personal experience is very limited.
ReplyDeleteAnd by "geopolitical situation" you mean refugees problem? I think that the model proposed in the article is similar to the one mentioned on last Advanced Simulations lecture - the one considering racial segregation, but it has more stress on clustering (grouping with members of own party).
DeleteSimulating crowd behavior is useful for architects and those responsible for the organization of mass events. You can not foresee everything, we are talking also about the behavior of "hardcore".
ReplyDeleteI never participated in a bar brawl, a football match brawl, and riots :). I do not know a similar software.
Yes, but to simulate mass events you have to take into account some architectural conditions. This model is way to simplified.
DeleteThis is very advanced and probably not so modern subject like a computer science. How difficult it is you may watch in news television or when you try to watch and guide an children party. When you feel advance enough you may try to go to let say Syria for vacation and try to be active reporter. But seriously how can I model this kind simulation when in probably most of my life I watched this types of events from ten thousand feet. Try to go into a hospital and make a first recon.
ReplyDeleteI'm afraid I don't understand how going to the hospital can help in simulating riots. Maybe you mean simulating riots' medical results? And modeling does not require experience in the subject, do you really think that people simulating, for example, atom bomb explosions have ever launched one?
DeleteHi Mikołaj, I think that simulating crowd behaviour is a very interesting topic. I try not to use violence to solve problems :) Of course, reading your post I thought about football matches. And of course, the last devastation of the stadium. I've never used this type of software, but I have to agree that this is a very interesting topic. I'm not sure whether this software will avoid dangerous situations, but I believe that there is something to fight for ;) On the other hand I have to agree with Kinga that this task seems very difficult to achieve.
ReplyDeleteCraft International, a company established by Chris Kyle (Americal Sniper) has a motto: "Despite what your momma told you, violence does solve problems." :) Don't you agree that sometimes it actually does?
DeleteThe simulation of crowd behaviour can be very useful in crowd control. More we know about crowd more we can do to avoid problems.
ReplyDeleteIn the 19th century, crowd behaviour was a significant area of study in the field of psychology. Early social psychologists suggested the origins of crowds in her instincts and mass imitation. Many theories have been propounded by various psychologists and sociologists to explain why a crowd behaves in a particular way. The behaviour of the crowd is always emotionally determined.
Nowdays the theory of crowd behaviour has moved away from the older perspective that viewed the individual as coming under the sway of the crowd and losing its capacity for rational judgement before the sweep of an overpowering emotional contagion. Instead, sociologists now explain crowd behaviour by the same sociological concepts that explain social group behaviour.
>> I realize that you cannot invoke riots to analyse crowd behaviour in stress situation, but what do you think about simulating crowd behaviour? <<
ReplyDeleteYou can actually learn from previous outbreaks and model the software after that. By simulating these events you can try out different strategies (both ones that worked in the past and the ones that failed) and observe what's happening. This is indispensable for law enforcement officers - there is no way to learn the result of your actions in the live outbreak.
>> What do you think about the behavioural model of the agents proposed in the article? <<
The statistical model is unfortunately the oversimplification of the reality. I think the model they used is way to simple.
>> Have you ever participated in a bar fight, football match brawl or maybe even riots? <<
Despite how I look and my passive aggressive behavior at times, I am not really aggressive. The answer is no.
Yes, I agree that the model is too simple.
DeleteAnd what do you mean by "passive aggressive behavior"?
Each simulation has to be connected with certain restrictions. It is not possible to involve all factors that affect on crowd behaviour. Considering that we have to focus on properly selected and balanced factors and their attributes. Otherwise we can receive totally far for reality results. Unfortunately I am not familiar with simulations yet, but this method looks fine. To be honest I think emotions of the day and let's say crowd "type" is main factor in this simulations (i.e. crowd of fire-fighters will extinguish fire, when in the same situation most of other groups of people will start panic and run far, far away). Of course after some simulations we can try to find out some kind of pattern of this behaviours. Simulations role is to show possible behaviours in situations which might be extremely difficult to reproduce in reality, for example Fire of huge building with concrete structures. We can compare this simulation only to similar event from the past. Human behaviours are connected with the trends, which evolve very dynamically.
ReplyDeleteRegarding last question can you share with us your experience ? :) (my answers are: bar fight NO, football match brawl NO or maybe even riots NO)
Finally! I thought nobody will ask :) Of course I consider bar brawls or football match brawls entertaining, but unfortunately I've never participated in proper riots.
DeleteI realize that you cannot invoke riots to analyse crowd behaviour in stress situation, but what do you think about simulating crowd behaviour?
ReplyDeleteIn my honest opinion it's too hard to simulate such behaviour. We have too many variables and we need too many assumptions to succeed. From what I know, we have tried to make such simulations with our national stadium and scientists have failed to do this.
What do you think about the behavioural model of the agents proposed in the article? Have you ever participated in a bar fight, football match brawl or maybe even riots?
I never had a chance to participate in a bar fight, I have seen a few of them. Most of the time they are connected with alcohol, so we can think about simulating drunken people behaviour. Good luck with that.
Do you think that some kind of human motion analysis software is able to detect possible conflicts or "hardcore" participants in the crowd? Do you know any software of that kind?
I think they can, I read about software that can identify potential terrorists on airports. It's still in alpha state, but they can find people with different temperatures, mimic and something like that. I'm wondering how efficient it is and how good it is. I had a chance to see a test on my own eyes and it wasn't accurate in my opinion. Predictions were far from being accurate.
Thaks for your answer and information about the software.
DeleteCan you tell it's name or paste a link to it's website?
The behaviour of people in crowd situations is a fascinating subject. All around the world people gather to pray, to celebrate some events, to demonstrate against rules or to fight. I reckon that all studies and research about behaviour of the crowd are very useful and deliberate. I hope more and more research will enable to avoid many tragic situations around the world when people die during common events or services because organizers don’t predict all mortal factors. I don’t know typical software to research the behaviour of people in crowd because this is my first meeting with modelling agent-based research. I think there are a lot of computer programs concerning modelling different people behaviours around the world.
ReplyDeleteI think that simulation can be useful but only in case if model is suited for some specific environment. Modeling crown behaviour in 100x100 meter square is in my opinion pointless.
DeleteThe simulation of the behaviour of crowds of people and swarms of animals (not just mythological ones) is also being applied to many other unusual situations, from designing better closed-circuit television (CCTV) security systems to managing the traffic of ships in harbours. The same technology has also been used to improve the understanding of archaeological ruins and to model entire ecosystems in order to design wildlife-management strategies.
ReplyDeleteThe technology can also be used to model animal behaviour. Massive is working with researchers at the University of Southern California (USC) School of Cinematic Arts and the Wrigley Institute for Environmental Studies, both in Los Angeles, to model animal behaviour on the conservation island of Santa Catalina, 20 miles off the coast. One aim of the research is to develop effective culling strategies for the island's bison population. These were introduced 85 years ago to make a silent movie and have since overrun the island and damaged local vegetation, say Eric Hanson, a professor at USC. The software will also be used to model other mammals and birds on the island.
As agent software becomes better able to capture complex real-world behaviour, other uses for it are sure to emerge. Indeed, this could soon become a crowded field.
Hey, thanks for bringing us this interesting article. Some time ago I've heard of a solution that can predict crowd behaviour and simulate how people would react during an evacuation. The researchers were comparing it with liquid physics and the flow of water. I can't remember the exact location I've seen it, maybe it was on Youtube or probably on Discovery channel back in the time I used to watch TV at all. I think that it is possible to detect specific motions that might indicate certain behaviours of individuals or even group reaction (i.e. people backing or trying to stay away from some location due to perceived danger). That is all very interesting and I am looking forward to see how the required technologies advance to achieve these results.
ReplyDeleteLiquid physics sound interesting. I've seen some videos on evacuation but I think that they simulate something more like promotion at Biedronka, not a real evacuation, where people trample each other and don't wait for their turn to exit, or enter the store :).
DeleteSimulating crowd behavior can be really great because you can predict how people can behave in some situation (ie evacuation or riots). I have never participate in bar fight or other riots. :-D. I think that systems can predict danger based on human posture and its movement and also how people are gathering in groups. I think that this is feasible and it's matter of time in which we can see this kind of systems.
ReplyDeleteI agree that this kind of system is possible to develop. I think that the main issue will be privacy - system of that kind will constantly monitor crowd and track people, for example in hypermarkets, malls or on the streets. Do you think that people will agree to be monitored for increased security?
DeleteSimulating crowd behaviour could one day be the strongest tool to avoid for example terrorism attacks. One person could start riots and the main task is to type that person from the crowd before anything happen. Today we can observe no borders politics. The consequence if this attitude is riots in many european streets. That system combined with city monitoring could preserve citizen against attack.
ReplyDeleteIn my opinion simulation cannot preserve attacks or riots, how do you think it is possible?
DeleteI think it is very unlikely for one person to start riots. Recent terrorist attacks in Europe were carried out by small groups of terrorists. And that increases chances of detecting at least one of possible attacker.
I realize that you cannot invoke riots to analyse crowd behaviour in stress situation, but what do you think about simulating crowd behaviour?
ReplyDeleteIn my opinion very interesting and usefull is simulating crowd behaviour in buildings. The simulation is a simplification of the real situation, but it can be succesfully usefull for having overall information how the crowd will behave in a dangerous situation. It could help to design a save, easy to leave, building with efficient escape routes. It also could be applied for designing e.g. streets in a newly built part of the city.
What do you think about the behavioural model of the agents proposed in the article? Have you ever participated in a bar fight, football match brawl or maybe even riots?
I have never been in such situation, so it is a bit difficult to compare the agents behaviour to the real situation.
Do you think that some kind of human motion analysis software is able to detect possible conflicts or "hardcore" participants in the crowd? Do you know any software of that kind?
Actually our Academy performes studies about unusual behaviours in a crowd, based on a vision analysis. I think the research is mostly carried out in Bytom. You can read a little about it in http://bytom.pja.edu.pl/laboratorium/laboratorium-hsl-wizji-komputerowej/.
What do you think about simulating crowd behaviour?
ReplyDeleteI think that it is a quite interesating issue. However, it is hard to simulate crowd behaviour. There are so many factors influencing every person. We can fail at the wrong assumptions.
What do you think about the behavioural model of the agents proposed in the article?
The model proposed in the study is oversimplified. Of course, it is a model of a reality but I thinkt that it is far too simple.
Have you ever participated in a bar fight, football match brawl or maybe even riots?
I have never participated in any bar fights, football match brawls or riots. I am afraid I don't know how people react confronted with such a situation.
Do you think that some kind of human motion analysis software is able to detect possible conflicts or "hardcore" participants in the crowd? Do you know any software of that kind?
I think it is feasible to detect possible conflicts. I think that some kind of a biological way is promising. A potential terrorist probabli will have different temperature, will sweat and so on.
We must remember that in terms of mass behaviour we are not so complicated as we want to think. Things that we can do are just variables in a very big and complicated equasion. Of course you can not predict every action and every implication of someone reaction but you can build an interesting and complex model that can predict most or possible situations. You wrote 'I think science should concentrate on software to predict and avoid conflicts.' - this is the reason of such software. If we understand such situation better we will be able to prepare better, predict and avoid :)
ReplyDeleteThe idea of simulating crowd behaviour isn't so bad when we want to plan for example emergency exits on a stadium. I don't have too much experience with simulations so far, but I'm sure that in case of crowds every model would be too simple. The point is to fit as many valuable parameters as possible for a particular scenario. I used to watch alive many football matches a when I was a bit younger, gained experience could be useful for making such a model.
ReplyDeleteSimulating crowd behavior is difficult, it depends on many factors. When you run a simulation you can easily miss something. I consider that the simulation can give answers to many difficult questions. It is an excellent tool for designers. Regarding the question: "Do you think that some kind of human motion analysis software is able to detect possible conflicts or" hardcore "Participants in the crowd?". I answer: analysis of speeds of movement, noise and a few other factors can give an alarm signal. I don't practice of amateur boxing:)
ReplyDeleteI realize that you cannot invoke riots to analyse crowd behaviour in stress situation, but what do you think about simulating crowd behaviour?
ReplyDelete> Ohh. You caught me here. I don't have any background on this matter and can't answer this question even though I did read the stuff from the links you have posted.
To be honest, I barely see the pursope of this kind of software. I think science should concentrate on software to predict and avoid conflicts. Do you think that some kind of human motion analysis software is able to detect possible conflicts or "hardcore" participants in the crowd? Do you know any software of that kind?
> No, I don't know anything in particular. I'm pretty sure that the software is not capable of understanding context as of today.
Oh that is very interesting problem – that is also problem of human nature – that is problem of what kind of simulation we could use to analisys human behavior in that kind of situation. That is no simple to create approximation to really human behavior. That is also another problem – that could be only simulation and we don’t have 100% relibility that we have ideal program to simulation. We could imagine about every situation and after that add that position to our porgram but we still don not have safety that our specification are exhaust. That is very difficult situation – but we have to think about it becouse it is important to people.
ReplyDeleteI don't know any of this kind of simulations but it sounds really interesting. I've just started to learn about social simulations hence I can't tell much about what is possible and what is not. This article motivated me to explore more about the subject. It would be pretty amazing to prevent conflict escalation by simulating human behavior.
ReplyDelete