- https://codevscovid19.devpost.com/
- To register
- https://github.com/csamuelsm/covid19-simulations
- simulation of the covid-19 dissemination made with Processing and inspired by the 'Why outbreaks like coronavirus spread exponentially, and how to “flatten the curve”' article on the Washington Post and written by Harry Stevens
- https://processing.org/download/
- Processing is a flexible software sketchbook and a language for learning how to code within the context of the visual arts. Since 2001, Processing has promoted software literacy within the visual arts and visual literacy within technology.
- https://www.nih.gov/health-information/coronavirus
- The current state of the art on covid research can be bout on
- https://www.supinfo.com/articles/single/4777-pymunk-moteur-physique-2d-vos-programmes-python
- Simulations in Python
- From the original simulation method
- Random walks
- Independence between spheres
- Infection by contact
- Change of direction after the contact (opposite direction)
- No smaller closed blocks (representing closed environment such as home / institutions etc)
- Our simulation method
- Dependence between certain spheres with some others
- Directed walks (ex: work) AND random walks (hobbies)
- BUT the walks are constrained by an environment (existence of « building » , creation of « roads »)
- Infection by being next to someone else (no contact necessary) and depending on a certain probability
- We could maybe implement a chance distribution with contact higher chance and then decreasing chance with increasing distance
- And I forgot to mention something important : in the base model, all spheres fully recover after a time t (based population = final population) but in our model we need to take into account the fact that people die (based population > final population).
- No change in direction after the infection
- Creation of a sort of hospital with a fixed capacity? (Need to think on how it should work)
- Disease growth rate curve displayed in the same time
- What we need to determine:
- The probability to get infected after a contact
- The probability of death
- Disease duration (at least an estimation)
- The time during which a person can infect someone else
- An average on how many people a patient infect during the first phase of the disease
- Our assumptions
- Focused on a specific system / country or worldwide? (Health systems are different between countries)
- The number of dependence between spheres (taking into account « family » , « work », « friends », « everyday routine activity »)
- We could maybe use graph theory to generate such kind of relationships
- Pros
- General simulation system that can be used later for another pandemic by changing some parameters
- More accurate
- Cons
- Computationally demanding ? (Need to create a map of size M, with a population of size P …)
- Population of 100 (can be changed) in a rectangle area
- Each collision has a probability of 50% to transmit the virus (can also be changed)
- There is a hospital in the middle (red circle) with a capacity of 20 patients (20% of the population, can be changed)
- We begin with a 20-years-old infected, the others have random ages
- When two people collide, they go away in opposites directions
- People goes to the hospital when the incubation period is finished and there is room for them
- Once in the hospital, people heal faster with a lower probability of dying
- In the initial population, there are 40 doctors that are able to go in the hospital. Other healthy people are not allowed to go inside.
- Dead people are still visible but don't collide, don't move and don't transmit the virus