A new approach for 'ab initio' synthesis of thin lens structure of optically compensated zoom lenses is reported. This is accomplished by an implementation of evolutionary programming that explores the available configuration space formed by powers of individual components and inter-component separations to obtain globally or quasiglobally optimum solutions for the problem. Normalization of the variables is carried out to get an insight on the optimum structures. The method has been successfully used to get thin lens structures of optically compensated zoom lens systems by suitable formulation of merit function of optimization. Investigations have been carried out on four-component zoom lens structures. Illustrative numerical results are presented. © 2011 Elsevier GmbH. All rights reserved.