This article presents a method to optimize concrete mix proportions with respect to different goals of economy and reliability or, equivalently, probability of failure. This method is based on a quadratic generalized ridge regression model to predict compressive strength of concrete for 28 days curing period and a linear regression model to predict cost of concrete. NSGA II is used to obtain reliable Pareto-optimal fronts with non-dominated solutions for different compressive strength requirements. Pareto-optimal fronts evolved by varying compressive strength requirements and probability of failure are analyzed. It is found that there is a nominal rise in cost as probability of failure decreases up to a certain limit for a given compressive strength requirement. However, there is a sharp rise in cost of concrete below that limit.