michael-iuzzolino / flappybio-neat Goto Github PK

Comparison values seem too low and speciation isn't happening when it should be.

Have to account for interspecies mating possibility (0.01 chance)

Too many organisms are created per species as speciation begins to happen. Total population should be constant (e.g., 150); however, with speciation, some species can have > 150 organisms, when 150 organisms should be the total number allowed in the entire species' population.

The species need to be culled correctly. If after 15 rounds there is no improvement, then only the top 2 organisms of the species should mate and produce enough progeny to refill the species.

Should the weights actually be bound between [0, 1], or [0, inf) ? Negative numbers may be skewing the similarity measure during checking genomes for speciation.

Must ensure that if same structures evolve at the same time within one generation, that multiple innovation numbers do not apply to the same innovation gene.

Need to create a list of innovations per generation replication, and check genes with same structure but different innovation numbers, and change them to be the same innovation number.

Species ID counter on the pygame screen is overlapping with double digits.