This dissertation investigates the value of deep reinforcement learning (DRL) within an agent-based model (ABM) of a large open-air venue. The intent is to reduce civilian casualties in an active shooting incident (ASI). There has been a steady increase of ASIs in the United States of America for over 20 years, and some of the most casualty-producing events have been in open spaces and open-air venues. More research should be conducted within the field to help discover policies that can mitigate the threat of a shooter in extremis. This study uses the concept of dynamic signage, controlled by a DRL policy, to guide civilians away from the threat and toward a safe exit in the modeled environment. It was found that a well-trained DRL policy can significantly reduce civilian casualties as compared to baseline scenarios. Further, the DRL policy can assist decision makers in determining how many signs to use in an environment and where to place them. Finally, research using DRL in the ASI space can yield systems and policies that will help reduce the impact of active shooters during an incident.

Background: The relevance of conducting this thesis is evident by the fact that bus and coach casualties have been “stubbornly stable” in Europe recent years and a need for investigating if a similar trend could be found in Sweden is therefore obvious. It was also important to add new knowledge to the bus and coach research in Sweden, since many areas were scarcely addressed. Aims: To describe bus and coach occupants’ injuries, crash and injury mechanisms generated in a traffic environment based on data from the medical sector. Additional aims were to investigate the injury reducing effect of a 3-point belt, the effect of cross-winds, and crucial factors in the emergency- and rescue response. Material and methods: Injury data analyses were based on a complete ten-year medical data set from a catchment-area with about 130,000 inhabitants. A number of crash studies with the scope in different crash phases were conducted by applying and elaborating the Haddon matrix as a framework. An additional framework, Protocol for Major Incidents was used in order to investi-gate the emergency- and rescue response to a severe coach crash. Results: Between the first and second five-year period, the incidence of injured in non-crash in-cidents was increased by 24%. In non-crash incidents, 54% were injured; 2/3 while alighting from a bus or coach. The pre-crash factor cross-wind, in addition to vehicle design, vehicle speed and road friction, was investigated in ten crashes. It was confirmed that cross-wind, in relation to vehicle speed and slippery road conditions, needs more attention. The importance of goods load-ing and passengers’ position in the bus, was indicated by the fact that a displacement of the cen-tre of mass rearwards with 10% increased the necessary coefficient of friction with, on average 45%, which in many cases corresponded to dry road conditions. Three Swedish rollover crashes were analysed with regard to the injury outcome, mechanisms and the possible injury reduction for occupants using a safety belt. A considerable increase in safety for occupants belted with 3-point belts was shown through limiting interior contacts, occupant interaction and the possibility of ejection. Crucial post-crash factors in the emergency- and rescue response showed that ordi-nary ways of working and equipment are not always useful and proper equipment for lifting a coach body is essential in the case of a rollover. Finally, the communication between the hospitals is important, and the telephone systems may be overloaded by calls from worried relatives and media. Conclusions: In non-crash events: Non-crash events constitute a majority of all bus and coach casualties with a high proportion of elderly female occupants among the MAIS 2+ injury cases. Boarding and, especially alighting causes many injuries to the lower extremities. In the pre-crash phase: Cross-winds do affect the safety of buses and coaches and requires more at-tention. Seat belt usage among bus and coach occupants has to be increased. In the crash phase: Rollover and ejection are the major causes behind serious and fatal injuries to bus and coach occupants, consequently, retentive glazing, pillars or rails need more attention. An upgrade from 2-point seat belts to 3-point seat belts yields an increase in the estimated injury re-duction from approximately 50% up to 80% for the MAIS 2+ casualties in a rollover crash. In the post-crash phase: In order to be able to lift a coach body proper equipment originated from experience and development is essential in a rescue operation of a crashed bus or coach. Fur-thermore, to improve the emergency response inside crashed coaches proper methods originated from experience need to be developed. Euro NBAP: Based on the results and conclusions generated in this thesis, a European New Bus and Coach Assessment Programme is suggested, which would provide bus and coach occupants with a assessment programme similar to the Euro NCAP.