This paper presents the modified global optimization co-algorithm based on the Mind Evolutionary Computation (MEC) algorithm for optimizing optical system designs. This kind of systems requires high precision. As a result optimization algorithms tend to converge slowly to a local optimum while trying to guarantee high quality of solutions. The concept of a co-algorithm helps to overcome this issue by identifying the most promising search areas and allocating there more computational resources. The outline of the proposed co-algorithm as well as its software implementation are described in the paper. The algorithm was utilized to optimize the structure of an optical Cooke triplet lens and helped identify various designs with best optical properties. All results are presented in the paper.