# Simulation Documentation
###### tags: `others`

[TOC]

## Finite-state Machine

digraph Emodel {
    node [shape=box] S, F, R, E, I, D
    S -> R
    S -> F
    S -> E
    F -> E
    E -> I
    I -> R
    I -> D
    I -> F
    
    {rank = "same"; S; F}
    {rank = "same"; R; D}
    
    rankdir = "LR"
}

### State
- $S,F,E,I,R,D$ or more
- a tmp queue, update to the state at the end of a period

### Expiring state
- $E,I$
- maintain a *heap* where key is the expiration period/day
- `while(top() is expird) pop()`
- iterating
- $I$

### Infectious State
- $I$

### Susceptible State
- $S,F$
- maintain $A$ balanced trees for each contact group
- $A$ is the number of age groups
- support `insert`, `remove`
- support *randomly choose $k$ elements*

## Simulation

### simulation of a period

def simulation_period(day, period):
    for U in infectious_states:
        for V in susceptible_state:
            U.infect(V, period)
    
    for U in expiring_states:
        U.expire()
        
    vaccination()
    
    for U in states:
        U.end_of_period()

### infect

def infect(V, period): # U.infect(V, period)
    for u in U:
        for g in group(u, period):
            sieve_algorithm(u, g, V, period)

### expiring

def expire(): # U.expire()
    while u = U.heap.top() is expired:
        U.remove(u)
        U.leave(u) # choose which state u is gonna be in

### at the end of a period

def end_of_period(): # U.end_of_period()
    for u in U.new_nodes_added_in_this_period:
        U.add_new_node(u)

## Random

- $U(0,1)$
- randomly choose node
- $\mbox{Exp}(\lambda)$
- determine period

## I/O
- one file discribe the whole graph
- input and output in the same format

## Implementation
- `cpp`
- SIR model with naive algorithm

## Scalability

- population
- days and periods to simulate
- size of age group
- contact matrix
- finite-state machine
- $I_{hosp}, I_{presym}$
- infecting vaccined group in lower rate