Modeling of nuclide densities as a function of time within magnetic confinement fusion devices such as the JET, ITER, and proposed DEMO tokamaks is performed using Monte Carlo transport codes coupled with a Bateman equation solver. The generation of reaction rates occurs through either pointwise interpolation of energy-dependent tracked particle data with nuclear data or multigroup (MG) convolution of binned fluxes with binned cross sections. The MG approach benefits from decreased computational expense and data portability, but introduces errors through effects such as self-shielding. Depending on the MG structure and nuclear data used, this method can introduce unacceptable errors without warning. We present a MG optimization method that utilizes a modified particle swarm algorithm to generate seed solutions for a nonstochastic string-tightening algorithm. This procedure has been used with a semihomogenized one-dimensional DEMO-like reactor design to produce an optimized energy group structure for tritium breeding. In this example, the errors introduced by the Vitamin-J 175 MG are reduced by two orders of magnitude in the optimized group structure.