A solver is used over a simulator to simulate Tokamak devices to look for the possibility to use them in aneutronic fusions in the combustion chamber of the Miranda reactors.
In order to do that a giant magnetic field must be used to confine 550keV particles.
In order to simulate thousand configurations, it is used 4 threads over C++ with improvements like using elliptic integrals to increase the simulation speed
The reactor cross-section divided by collision probability is too high so a lot of simulations must be run to increase the performance.
It is improved the Aeff that is the averaged reactor section by using the “23 fellow system”, which involved to generate vectors of parameters with a variation between them using genetic algorithm but when having more than 23 vectors the worst of them is erased, then the percentage of variation is reduced every time a new vector is obtained as a variation of the 23 fellows.
After 10 kilosimulations of different Tokamaks structures with 4 to 12 toroidal coils are been simulated. Thanks to the 23 fellow system the performance is increased in very few simulations as can be seen in the yield table using logarithmic scales:
