Case Study on York University
Research and development teams have been increasing their efforts to find solutions to combat SARS-CoV-2. Historically mass vaccination has been an important aspect of virus management and immunization and so it was believed that vaccines would be the best solution to ending the pandemic. One way this was successfully done in the past was through drive-through facilities as seen in mass influenza vaccine’s. However, additional measures are needed when adapting them for usage in COVID-19 vaccination centres due to the need for safe physical distancing.
Ali Asgary and Mahdi M. Najafabadi in collaboration with Richard Karsseboom and Jianhong Wu of York University produced an excellent and interesting paper on how they used AnyLogic to produce a simulation so that they could enhance the planning, design, operation and feasibility and effectiveness of facilities such as the drive-through mass vaccination. They also produced a webinar on discussing the model on the AnyLogic website.
The model itself incorporates a hybrid approach by integrating both discrete event and agent based modelling techniques. Within the model, there is a physical layout of the facility in which several agents interact with each other based on specifically defined logic as well as implementing policies that are passed to the model through user input. This is helpful in the simulation as it allows users to customise parameters such as, different numbers of staff, drive-through lanes, screening, registration, immunization and lastly, recovery times. These are all important parameters when it comes to the optimisation of mass vaccine drive through’s and will allow for them to be tested to be used efficiently in COVID-19 and any potential future pandemics.
The layout of the drive-through consists of ten initial service lanes however they can be turned off as a user input. Cars enter the model at a given rate per minute which can also be defined before run-time, before going through a screening booth which is a single service station in the model that then sends the cars into different lanes. Some cars will get rejected at this stage and take a bypass lane that causes them to exit the model and the rest of them will be spread across the service lanes based on existing queues. Within the service lanes are multiple service stations, these offer registration and consent, vaccination delivery and finally, recovery. When running the base model results showed that most cars spend between 80 to 90 minutes in the drive-through on average however the earlier the cars came in the shorter their times were. The more cars increased queue times and the maximum time spent in the drive-through was 96 minutes. Multiple tests were run, and parameters were changed to find the efficiency of the drive-through when changing the number of lanes for example. When only one lane was open the average time spent was close to 180 minutes but when 10 were open the average was around 75 minutes. As expected, more lanes increased efficiency which is useful to know when designing a facility like this.