This paper presents an inverse problem in computed tomography using a simplified simulator operating in a two-dimensional space and proposes methods for solving it. The solutions discussed are based on a metaheuristic algorithm—the Group Teaching Optimization Algorithm (GTOA)—as well as a hybrid approach that combines this metaheuristic with two deterministic algorithms: Nelder-Mead (NM) and Hooke-Jeeves (HJ). The study includes performance evaluations of each algorithm, accompanied by graphical representations and a direct comparison of fitness function values and computation times for each solution. The results demonstrate that the hybrid approach consistently produced solutions of comparable or superior quality, with one hybrid instance significantly outperforming the standalone GTOA. Additionally, the hybrid algorithms required substantially less computation time, ranging from 32% to 85% of the time taken by GTOA.