Abstract: This paper presents an improved genetic algorithm (GA) tominimize the weight of a truss with discrete sizing, continuous shape variables.Because of the nature of discrete and continuous variables, mixed codingschemes are proposed, including binary and float coding, integer and floatcoding. Surrogate reproduction is developed to select good individuals tomating pool on the basis of constraint and fitness values, taking fullaccount of the character of the constrained optimization. This paper proposes anew strategy of creating next population by competing between parent andoffspring populations based on constraint and fitness values; so that thelifetime of the excellent gene is prolonged. Standard examples are solved,numerical solutions are shown to be better than those in the literature.