For the safety and security of our communities, it is paramount that both people and authorities be prepared for public emergencies and mass evacuations of crowds. Planning practices, however, need not be confined to mathematical models and architectural design of the infrastructure. Rather, we should recognise that the preparedness of individuals, their knowledge and the effectiveness of their responses could make a significant difference in increasing their chance of survival. Modern evacuation planning practices should move on from viewing people as the problem to control and should rather seek to harness the potential and critical role of the public in disaster mitigation. Put differently, the public should be regarded as an ally and become part of the disaster management solutions. This can be achieved through education and peace-time training, and it could save lives in times of crises. Pursuing this approach, however, requires that (i) researchers enhance their knowledge of optimum individual evacuation strategies and actions and (ii) the public be educated and equipped with evidence-based and scientifically-proven knowledge and training to deploy during crises. The notion of behavioural intervention and public education/training could also be potentially extended and adopted as a pragmatic method of enhancing community preparedness and response to crises other than mass crowd evacuations. This could include health emergencies and also wildfire evacuations.
Mass emergencies have become part and parcel of living in dense urban communities. They are rare incidents but could have catastrophic consequences if not dealt with and planned for competently. None of us is a stranger anymore with the news of terror attacks, mass shootings or building fires that occupy the social media feed every now and then plus the more recent bushfires in Australia and the ongoing coronavirus hysteria. Such incidents of mass emergency have become so ubiquitous that singling out one incident in this note might be unnecessary as it may downplay other similar tragedies that have taken lives and caused physical and mental trauma to people around the world.
Perpetrators of evil acts who mean to harm the public, such as terrorists or mass shooters, often target high-density spaces. Crowded places such as airports, theatres, sport stadiums, nightclubs, festivals, places of worship, high-rise buildings or shopping centres, are soft targets. In general, any public space where there is a high concentration of people could be vulnerable as it allows those with nefarious intentions to maximise the catastrophic impact of their actions. But our problem at hand is not confined to the cases of intentional man-made disasters. Public spaces and buildings are also often vulnerable to cases of fire or earthquake emergencies. Regardless of the nature of these emergencies, an important issue here is that our urban communities need to be prepared for incidents which require a high concentration of people, a mass crowd, be swiftly evacuated. In such instances of acute emergencies, every second counts as it can spell the difference between death and survival. Therefore, planning for mass evacuations of crowds is an essential and paramount component of disaster preparedness for our growing urban societies.
Given the significance of the problem that was described and its implications for saving lives in times of emergency, an ample amount of research has been conducted, mostly within the last twenty years or so. The aim has been to better understand evacuation processes and to ultimately equip authorities and practitioners with tools and solutions that can be used for evacuation preparedness. A major cornerstone of the research in this area has been the development of numerical prediction models for evacuations. Such computational tools have advanced substantially and have become increasingly sophisticated, from models that drew analogies between the movement of crowds and that of fluids and replicated crowd flows based on the principles of fluid mechanics—the so-called macroscopic models—to the modern agent-based models that recognise the composition of crowds from individual entities. Such microscopic models are, in fact, often reminiscent of the traffic simulation models that we use for transport planning. In fact, a common principle applies here: one can run these computational tools, simulate the movement process (of pedestrians or vehicles), test various likely scenarios, and obtain estimates of the system performance, in this case an estimate of the evacuation time. This is, in and of itself, immensely useful. But does having an estimate of the likely evacuation time per se translate to shorter evacuation times and saving lives? The answer is likely negative as this would be just the first step.
To answer this question, researchers have predominantly resorted to numerical computational models. The problem is often referred to, in the scholarly literature, as evacuation optimisation. An inspection of the literature shows two main streams of optimisation approaches in this domain. One stream treats the problem as mathematical optimisation programming. In formulating such mathematical models, by using the numerical simulation tools, researchers aim to identify evacuation plans that maximise system performance. The solution to such optimisation problems often determine a path and/or departure-schedule plan for the entire crowd that minimises the total time to evacuate compared to all other path and/or departure-schedule schemes. The other stream of optimisation research also often resorts to numerical simulation models but views the problem from the perspective of infrastructure design. Researchers in that domain have investigated architectural designs that best support evacuation processes and have identified a range of solutions that are assumed to facilitate the movements of crowds.
The scientific value of these mathematical and architectural methods in evacuation planning cannot be ruled out. However, there are reasons why neither of these two approaches have not made their way effectively to the practical domain. When it comes to mathematical models, one major issue is that the optimised solutions often do not come with clear methods of implementation. In fact, enforcement of optimised path and/or departure-schedule planning solutions often requires that an entire crowd of occupants be guided by a central body. Such central authority, however, does not often exist during evacuations. Moreover, these mathematical models only address two specific aspects of evacuation response: path choice and/or departure schedule. Whereas, evacuation response is comprised of a multitude of dimensions. Local microscopic aspects of individual evacuation behaviour are often not addressed within this approach. In regard to the design of the infrastructure and its relation to evacuation planning, two major issues need to be considered. First, most of our existing facilities and buildings have not been designed optimally to support mass evacuations and making them suitable for evacuations, even if we assume that we have the perfect scientifically proven design solutions at hand, may require major alterations to the design of current buildings. This is not in many cases practical. On the other hand, the effectiveness of many of the design solutions are currently in dispute in the scholarly literature. We should note that the majority of the design solutions have been obtained from computational models, and these models, like any other prediction tool, are never perfect. They often suffer from issues of modelling artefacts. Recent developments in the experimental domain of crowd research has succeeded in testing some of these design hypotheses and, in many cases, the experimental observations have failed to validate recommendations of the numerical models. There is a perfect example to be made here, a well-known design solution that is perhaps no longer confined to the research domain and is even ubiquitously believed by lay persons as a potential effective way for crowd management. For a long time, there has been this counterintuitive assumption that by partially obstructing the area in front of an exit one could facilitate the flow of people. As a researcher who has been active in this domain for several year now, it has been a repeated personal experience to receive this comment after presenting research outcomes to an audience who may not even have expertise in this field: “it is common knowledge how you can accelerate crowd flows, by placing blocks in front of exits, maybe in an optimal way though”. To many transport scholars, this may be reminiscent of the so-called Braess Paradox in road network design. But the analogy may not hold below the surface. Regardless of the fact that in many existing buildings, from an architectural and aesthetic perspective, it is rather unimaginable to place permanent blocks in front of exits, recent experimental testings have shown that this solution is not even effective and may actually have consequences contrary to what we believed. Evidence is gradually emerging to suggest that such drastic solutions that may seem counterintuitive at first glance, may in fact be counterproductive .
Given the challenges of conventional evacuation optimisation approaches that were laid out in the previous lines, the question is whether there are more practical ways for evacuation management, potential solutions that are overlooked. The answer is affirmative. The contemporary research in the field of crowd dynamics is producing evidence that suggest there is considerable benefit to be gained through modifying or influencing the behaviour of individual evacuees during mass emergencies. This approach that I refer to as behavioural optimisation, or better said, behavioural intervention, is in recognition of the fact that the efficiency of an evacuation process is the collective outcome of the strategies that individuals choose. Hence, improving individual strategies could majorly benefit a system of evacuees. This, however, is itself a highly nuanced and multi-dimensional problem as evacuation response is comprised of several behavioural layers and types of decision-making. It would be the researchers’ task to identify behavioural layers that could be improved and to discover how they should be improved in order to increase collective efficiency. In tackling this question, parametric numerical models could be handy here too as they can provide an inexpensive behavioural laboratory to test various behavioural strategies and discover the optimum behaviour . Re-emphasising on our previous caveat about the possible artefacts of the numerical models, once informed hypotheses were made, one would need to put them under experimental scrutiny to ascertain their effectiveness before educating the public. This approach of behavioural intervention and enhancing individual preparedness is assumed to offer solutions to the practical challenges of the more conventional methods in mass evacuation planning, and deserves to be explored more closely.
Traditional approaches in crowd evacuation management have not typically viewed the public as a potential ally. Rather, the conventional view of panicking irrational crowds that depicts people involved in crises as non-thinking individuals has hindered such potential efforts. While a detailed discussion on the suitability of the panic theory and its contribution to crowd management practices is beyond the scope of this note , it may suffice to say that recent developments in social psychology have challenged this perspective . It has been shown that even under acute stress, people are capable of making decisions and recalling their training to varying degrees [5, 6]. We do recognise the importance of training and preparedness for first responders, but why not think the same way for the zero responders, i.e., the public . Such public training programs could embody a wide range of individual actions from rendering first-aid medical assistance to those injured—that could immensely contribute to the mitigation of fatalities— to choosing efficient self-evacuation strategies that could help the person herself as well as others in the scene. In theory, the potential benefits of the behavioural intervention method could also be utilised in planning for public crises of various natures – other than mass crowd evacuations – including public health emergencies and bushfire evacuations. For example, a study has shown that non-pharmaceutical interventions intended to reduce infectious contacts between persons during the 1918 Influenza pandemic – or, what we now call social distancing in the face of the recent global pandemic – could potentially reduce death rates to nearly 50%, according to the historical data from the United States . This could be only one proven example and recognised dimension of how behavioural intervention could be used as an effective and pragmatic tool in managing public emergencies.
One should note that when a crisis strikes, there is always a time window when individuals on the ground are on their own, and when no central authority has taken control to guide and assist them. Within that period of time, i.e. the silent gap, it is the action and preparedness of the individuals embroiled in the emergency that has the most significant impact on their survival. Therefore, modern emergency planning practices should recognise the role of the public and their awareness and preparedness in disaster mitigation. Rather than placing blocks in their way which may even hinder their survival, we should consider educating the public and equipping them with the knowledge of the best evacuation strategies under various kinds of mass emergencies. How these education or training programs should be delivered, however, is another matter to be addressed by the researchers.
 M. Haghani, E. Cristiani, N.W.F. Bode, M. Boltes, A. Corbetta, Panic, Irrationality, and Herding: Three Ambiguous Terms in Crowd Dynamics Research, Journal of Advanced Transportation, 2019 (2019) 58.
 R. Lovreglio, V. Gonzalez, Z. Feng, R. Amor, M. Spearpoint, J. Thomas, M. Trotter, R. Sacks, Prototyping virtual reality serious games for building earthquake preparedness: The Auckland City Hospital case study, Advanced Engineering Informatics, 38 (2018) 670-682.
 M. Kinateder, P. Pauli, M. Müller, J. Krieger, F. Heimbecher, I. Rönnau, U. Bergerhausen, G. Vollmann, P. Vogt, A. Mühlberger, Human behaviour in severe tunnel accidents: Effects of information and behavioural training, Transportation Research Part F: Traffic Psychology and Behaviour, 17 (2013) 20-32.