This dataset contains the a Matlab version of the BESOT algorithm, a bi-level evolutionary method for size optimization of two-dimensionall trusses. The aim of the algorithm is to find the lightest structure whose members satisfy stress constraints. Given a fixed volume, the BESOT algorithm works on the ground-structure of the truss, increasing and decreasing the cross-sectional areas of the members according to a sensitivity measure based on their stresses. Once a set of stress constraints is satisfied, the volume of the structure is reduced and the member areas are rebalanced. On the other hand, if the constraints cannot be satisfied, the total volume is increased. The algorithm stops when the gap between the largest infactible volume and the smallest feasible volume is negligible. To illustrate the behavior of the algorithm, we also provide the files for eight problems, including four cantilever structures and four arch-shaped trusses. These files can be modified to generate a wide variety of trusses, even allowing some members of the structure to have a fixed area.