This work addresses the traffic network design problem when day-to-day uncertainties in travel demand and link capacity are taken into account. Specifically, this work proposes a network design formulation that uses a strategic behavior approach in which total demand and link capacity are treated as random variables, and a strategic user equilibrium results in fixed equilibrium link proportions. The bilevel model is formulated, system performance metrics are derived, and a solution method is then developed according to a tailored genetic algorithm. Results under varying levels of volatility reflect possible suhoptimal project selection when a deterministic modeling approach is used.