I set out a program containing a genetic algorithm I wanted to test to search R(5,5) = 43 for a counter example.
The algorithm rates each graph with a fitness value which is equal to the total number of 5 clique sub graphs.
After 3 days of running I got down from average values of 1,400 to 2,500 to a lowest value of 848 currently.
Does anyone know where I could find the lowest known fitness value for the specific problem of 5,5,43 that someone else may have found? (The goal is to find 0 to reach a valid counter example and create a new lower bound of 44.)
If someone has found lower values, it might be nice to take the chromosome for those graphs and plug them into my algorithm.