In this paper, a new methodology is presented to solve the tolerance allocation-process selection problem simultaneously The prob lem is modeled using discrete and continuous variables and is transformed into a model with only continuous variables by defining an efficient tolerance-cost curve for each component Since the efficient tolerance-cost curve is neither convex nor concave, nonlinear programming method ologies cannot be directly applied The tolerance-cost curve is piecewise linearly approximated and an efficient methodology is developed to solve the problem The method starts with a solution which minimizes the objective function value but is not feasible The infeasibility is iteratively reduced in a way that the increase in the objective function value is minimal Computational analysis indicates that the method is very robust and requires negligible CPU time