Steel plate girders are generally used to support heavy loads and span long distances. As these girders are made of steel plates, they can be custom designed to suit consumer’s requirements. The designer has the option of selecting suitable dimensions for the flanges and web thereby facilitating smaller flange area with deep webs. Further, the stability of these girders against web buckling can be enhanced by providing stiffeners or a corrugated web. These positive effects of using plate girder can be enhanced by using optimum design, so that cost effective and structurally practicable girders can be used in construction. It is noteworthy that studies on optimization of plate girders are scarce and therefore the focus of the present study is to develop genetic algorithm (GA) based formulations using MATLAB for the optimum design of trapezoidally corrugated web plate girders (CWPG). Parametric studies are also carried out to identify the influence of design variables on optimum design. Finally, an equation for estimating the weight of an optimized girder is successfully developed.