Abstract
Genetic algorithms that can solve problems in conduction heat transfer have in the past encountered difficulties with increasing mesh resolution. The reasons for these difficulties were identified to be due to the np-hard nature of the problem. They were successfully resolved by means of a novel "genetic engineering" operator. As the data show, the new algorithm displays good convergence and can solve much finer meshes than previously possible. In addition, the algorithm was tested on different boundary conditions, such as fixed temperature, convection, and prescribed heat flux. Finally, the effect of different parameters and operators on the accuracy of the results was studied.