Faster-is-slower effect in escaping ants revisited: Ants do not behave like humans
Introduction
It is already known that under normal conditions, ant and human fundamental diagrams (FD) are different (John et al., 2009). While the former displays a constant velocity for all densities, the pedestrian and vehicular FDs always show a monotonically decreasing velocity for increasing density (see for example Seyfried et al., 2005). Moreover, ants do not produce jamming (Dussutour et al., 2004, John et al., 2009).
Contrary to what the title of Soria et al.’s paper (2012) would suggest, the authors claimed that there are differences between ants and humans in highly competitive situations such as emergency evacuation through narrow exits. They reported the observation of the faster-is-slower (FIS) effect in escaping ants stressed with a chemical repellent at different concentrations. Even though the FIS effect has been reported for simulated pedestrians via de social force model (SFM) (Helbing et al., 2000), it is not enough for justifying an analogy between ants and humans when egressing through a narrow door. One should not be misled by the title of the paper since the authors clearly state that the mechanisms causing the FIS effect in ants are not the same as those in the SFM simulations. So, although the “FIS effect” was reported in this paper, using it as a proof that ant and human egress is similar, it is not correct.
During different types of emergencies people can adopt different behaviors depending on the demand and capacity of the means of egress. The balance between the demand and capacity is given by several factors such as the kind of physical threat, information and subjective perception of danger, the number of people and the widths of the means of egress. As long as the physical threat is not imminent or not directly perceived (for example an alarm, but no smoke or fire), people tend to be cooperative (Kretz, 2010). The shorter the time (or the smaller the width of a door or stairway) available for a safe egress is, the lower the degree of cooperativeness results. As there is less time available for escape from danger, decisions under stress could be taken (Keinan et al., 1987), which could result, for example, in choosing the main entrance instead of the nearest exit as means of egress. In the extreme case that the time available is very scarce to escape from a sure death, the predominant behavior would be the individual self-preservation.
In such a situation, people could choose rushing or not rushing toward the exit. As this decision has an impact on the payoffs of each agent and the whole group, it can be studied from the point of view of game theory. Heliövaara et al. (2013) have shown that jamming and clogging may be caused by people acting rationally, even when this rational individual behavior results in a bad strategy for the group.
An example of such egressing behavior, which saturates the capacity of the egress door, is the fire at “The Station Night Club” (Rhode Island, USA, 20 February 2003), where an amateur camera recorded the tragedy (http://www.youtube.com/watch?v=OOzfq9Egxeo). There, it can be seen that, at a given moment, most of the people tried to egress simultaneously through the main door, causing the blockage of that door (Fahy et al., 2012).
The behavior of rushing toward a door was also observed in animals in less frightening situations.
Saloma et al. (2003) found this response when studying the egress of mice from a water pool.
Zuriguel et al. (2013) also observed the same response when studying the anxious passage of sheep through a narrow door when they were to be fed.
Also, this selfish evacuation behavior is the one assumed in the paper where the FIS effect was first reported (Helbing et al., 2000). The cause of this effect is the high tangential friction between particles in contact (Parisi and Dorso, 2007). Moreover, the FIS effect was recently verified experimentally in granular media (Gago et al., 2013), herd of sheep (Zuriguel et al., 2013), and humans (Garcimartín et al., 2014). In all cases the FIS effect appears due to jamming and clogging at the door, producing high frictional forces.
On the other hand, the FIS effect reported for ants (Soria et al., 2012) was not caused by any kind of frictional contact, jamming or clogging. This behavior was also confirmed in another experiment with Argentinean ants stressed with temperature (Boari et al., 2013) in which, contrary to the FIS effect, the “faster-is-faster” effect was found, even when ants were close to dying by temperature (if it had risen a little bit further).
The present work is based on the video recording from the experiments performed in Soria et al. (2012). Here we used image processing technics for obtaining the individual trajectory of ants. From this information, velocities and densities can also be studied. These data allowed us to demonstrate the claim that ants do not jam nor clog near the exit and thus that the FIS observed has no relation to the FIS effect in other animals’ systems. To be even more specific, we are going to compare the more relevant metrics obtained from ant data with the corresponding ones from simulations with the SFM producing the real FIS effect. As a consequence, it will be evident that the FIS effect in ants is not the same as in other systems relevant to the area of highly competitive egress.
Section snippets
Materials and methods
We analyzed the recorded video of the experiments reported in Soria et al. (2012) when studying the egress of Camponotus mus (Roger) ants stressed with aversive stimuli through a narrow exit. A detailed description of the experiments can be found in that paper; here we only summarize the main features.
Approximately one hundred ants were placed in a transparent arena consisting of a floor, walls and a ceiling so high that ants could not get on one another, thus maintaining the system
Distribution of time lapses “dt”
In this section we will confirm the FIS effect, reported in Soria et al. (2012), using different metrics and study the distribution of dt.
By grouping the time lapses dt (see Section 2.1.1) according to the citronella concentration used in the trial, the complementary cumulative distribution function (CDF) of dt can be computed for the four distributions as shown in Fig. 2B. The lowest distribution corresponds to 75% citronella concentration, confirming the reported result when the mean
Discussion
The fact shown in the present work and in Boari et al. (2013) that ants do not rush toward the door in a selfish evacuation behavior when stressed with citronella or heat (endangering their own individual life) can be because social insects (ants and honeybees, among others) are not expected to behave similarly to other animals because they have a different life cycle than that of individual animals (including humans), as the reproductive unit is not the individual but the colony. Ant
Conclusions
In the present work we have provided new data extracted from previous experiments by means of image processing technics. The data consisted of the positions of ants as a function of time (trajectories) during the egress under stressed conditions, which allowed us to study velocities and densities at different locations inside the arena.
The results from the ant experiments have shown that ants distributed uniformly over the available area during the egress process and did not generate high
Acknowledgments
Roxana Josens and Daniel R. Parisi are Scientific Researcher at CONICET (Argentina). This work was supported by Grant PICT 2011-1238 (ANPCyT, Argentina).
References (23)
- et al.
Efficient egress of escaping ants stressed with temperature
PloS one
(2013) Ecological consequences of traffic organisation in ant societies
Phys. A: Stat. Mech. Appl.
(2006)- et al.
Power-law distributions in empirical data
SIAM Rev.
(2009) - et al.
Collective decision-making and foraging patterns in ants and honeybees
Adv. Insect Physiol.
(2008) - et al.
Optimal traffic organization in ants under crowded conditions
Nature
(2004) - et al.
Amplification of individual preferences in a social context: the case of wall-following in ants
Proc. R. Soc. B
(2005) - et al.
Panic or not in fire: clarifying the misconception
Fire Mater.
(2012) - et al.
Ant traffic rules
J. Exp. Biol.
(2010) - et al.
‘Faster is slower’ effect in granular flows
- et al.
Experimental evidence of the “Faster Is Slower” effect
Transp. Res. Procedia
(2014)
Simulating dynamical features of escape panic
Nature
Cited by (53)
Incorporating genetic algorithm to optimise initial condition of pedestrian evacuation based on agent aggressiveness
2021, Physica A: Statistical Mechanics and its ApplicationsCitation Excerpt :Thus, the FFCA model has been used extensively in recent years to investigate pedestrian dynamics phenomena such as the anticipation floor field describing non-local anticipation behaviour [14], long-term congestion anticipation and aversion [15], pedestrian group behaviour [16], and the sub-mesh system enabling high-density crowd representation [17]. In the study of pedestrian evacuation dynamics, the compact of pedestrian aggressiveness is one of the most studied topics [2,3,8,18] [19–21]. The ‘faster-is-slower’ effect, which refers to the phenomenon of a statistical increase in total evacuation time as individuals push harder through a bottleneck, is suggested by Helbing et al. (2000b) based on simulation study [8], and extensively studied by both computational and empirical approaches [2].