Abstract
This article proposes a new algorithm for grouping problems that is a grouping version of league championship algorithm (GLCA). We compare the performance of GLCA with several well-known algorithms published in the present literature and select a set of 20 most widely used benchmarks of cell formation sample problems posing as a grouping problem. We used a truncated geometric algorithm to find the number of initial cells. Our computations reveal that GLCA can reach the best-known solution for 17 of the 20 benchmark problems, and improve the solution of three others with a 1.4% average gap.