Active asset-investment and portfolio-construction strategies for infrastructure systems are developed. Two primary questions are explored: how to allocate a limited budget among the assets to enhance resilience of an infrastructure portfolio against extreme events and how to decide the most appropriate time to invest in a portfolio of infrastructure assets under information uncertainty. A portfolio optimization model with the objective of minimizing the economic loss from extreme events under the budget resource constraint is developed. Because of the network effect of the assets in a large-scale portfolio, two types of algorithms are developed to improve the computational efficiency of portfolio selection: an algorithm that restructures the objective function as a monotonic non-increasing function with a Taylor series expansion and uses the first-order term as an approximation, which does not consider the effect of simultaneous investment in two or more assets, and a heuristic algorithm that systematically considers the network effect of the assets in the portfolio via consecutive iteration. The results show that the investment decisions given by the heuristic algorithm reduce the expected value of the economic loss given by the approximation algorithm, significantly increase the expected return on the investment portfolio, and improve the system resilience under extreme events. A real-options-based approach is used to determine the optimal investment time and investment criteria for a portfolio of infrastructure assets under information uncertainty. First, a single-option framework for portfolio investment under a single uncertainty is developed, and applied under different growth scenarios. The resulting investment criteria are compared to the net present value (NPV) break-even point, and illustrate the merits of the advanced real-options-based investment model in both the deterministic and stochastic cases. Theoretical considerations and real-life cases show why the NPV rule cannot yield the optimal investment decision in either scenario. The single-option model is then extended to a multi-option framework for a portfolio of interdependent infrastructure assets under multidimensional uncertainties, with the aim of deciding the selection of assets for investment and the optimal time to invest in each of them, that is, whether investment should be made immediately or postponed to maximize the return on the portfolio. The portfolio planning analysis begins with a static NPV framework to quantify the marginal investment payoff of interacting assets and then considers the value of a growth option contingent on the information uncertainty. An algorithm based on dynamic programming and least-squares Monte Carlo simulation to search for the optimal investment decision and the compound option value of the portfolio is proposed. The results show that the multi-option model increases the investment value of the portfolio in both the deterministic and stochastic cases by allowing flexible investment timelines for individual assets. The stochastic scenario further reveals the advantage of the multi-option model as volatility increases, and shows that the model could serve as an effective dynamic adaptive decision support tool for multi-period investment in balancing the return on a portfolio versus risk by incorporating new information as it becomes available.