The free convection in a divided enclosure is an ongoing attractive research topic due to its occurrence in various thermal applications, which can be utilized as a dual-purpose application to suppress or enhance the heat transfer rate. In this investigation, a vertical enclosure divided by adiabatic louvers is geometrically modified. To this end, circular perforations are made along the louvers. Then, the steady laminar free convection in the modified divided enclosure is experimentally investigated versus design parameters. The design parameters are the Rayleigh number (7 × 103 ≤ Ra ≤ 1.45 × 104), louver’s slant angle (0° ≤ φ ≤ 150°), perforation’s diameter to louver’s width ratio (0.2 ≤ d/w ≤ 0.6) as well as enclosure’s aspect ratio (7.8723 ≤ Ar ≤ 9.3671). After that, a mathematical correlation is extracted for the mean Nusselt number (Num) versus the design parameters. In the following, the differential evolution (DE) algorithm and wingsuit flying search (WFS) optimization algorithm are combined in the frame of a novel combinatorial optimization algorithm. The combinatorial algorithm is then employed to optimize the obtained mathematical correlation. It was reported that the maximum heat transfer rate corresponds to the highest level of the Rayleigh number, d/w ratio as well as Ar. Moreover, for the minimum heat transfer rate, the Rayleigh number, d/w ratio as well as Ar, must be at the lowest level. The maximum and minimum heat transfer rates both occur at the same critical value of the louver’s angle (φcrit ≈ 90°).