This is the project of one of the teams in the SIGEvo summer school 2018.
This site can be accessed also through the short url https://git.io/wasabitigers
The final presentation at the summer school is available from GDrive
A new evolutionary algorithm inspired from the tiger mosquito plague fighting techniques. The only things that change is the selection mechanism, which uses a pheromone to represent the direction in which search should go.
An algorithm that removes mosquitoes that carry dengue virus, zykova virus, yellow fever virus. You will choose a pheromone to eliminate mosquitoes, a good pheromone and a bad pheromone to make a pheromone that attracts tiger mosquitoes.
- Initialize the pheromone.
- The way to initialize the pheromone fitness is to use the NK_landscape method.
- A detailed code of nk_landscape is shown on this page.
- Measure the fitness of pheromones.
- After initialization, the fitness function uses PP (Pheromone-Potential).
- We will use two different update methods for the pheromone value (P) of each individual.
- Based on the whole inheritance:
if x_i selected: P(x_i) += constant_value P(parents[x_i]) += Q^h*P(x_i)
- Based on their children
if x_i selected: P(x_i) += constant_value + k * size(childrens)
- The contents of the PP are shown in This page.
- Use the tormant selection to get the results.
- We use the new selection scheme that the individual is given the score stochastically.
- The random value
r
is sampled from the uniform distribution. - Select the
N_tournament
solutions inX
at random. - If
r <= sr
, the best candidate solution is set asx* = max_x ( f(x_1), f(x_2), ..., f(x_N_tournament) )
otherwise, the best candidate solution is set asx* = max_x ( P(x_1), P(x_2), ..., P(x_N_tournament) )
. - Return the best solution
x*
- The random value
- A detailed code of selection is shown on this page
- We use the new selection scheme that the individual is given the score stochastically.
- Cross over each result to get a new generation and perform mutation.
- repeat
A more extended description will be shown in this page
5 students from all over the world, and a tutor.
The tigers team tries to model the spread of mosquitos.