"Increasing penetration of renewable energy in the power grid is creating new challenges for generation planning. High levels of renewable energy - primarily wind and solar - results in high levels of uncertainty and variability at time scales as fast as seconds to minutes. As a result, conventional generation needs to have adequate flexibility to maintain the balance between supply and demand at fast time scales. Therefore, while planning the future conventional generation mix, a problem known as generation expansion planning (GEP), it is important for power system planners to incorporate flexibility. In particular, it is valuable to include high temporal resolution operational details into GEP in order to ensure that conventional generators will have sufficient flexibility to be able to counteract the intra-hour fluctuations in renewable generation. The main challenge for incorporating intra-hour operational details into GEP is that for existing formulations the problem size scales dramatically as the temporal resolution increases, leading to prohibitively high computational cost. To address this challenge, this dissertation proposes a novel, computationally efficient GEP formulation which ensures the selected generation mix has sufficient capacity and intra-hour flexibility to balance the variations in renewable generation. The method proceeds by using historical data to develop time-independent bounds on ramp (MW/5 min) as a function of load and renewable generation (MW). The solution quality of the proposed approach is validated using the IEEE RTS test case by comparison with two existing GEP techniques - unit clustering and representative weeks - and by assessing its sensitivity to generation parameters. The method yields generation mixes with costs as low or lower than the alternatives, but with much faster run times and including higher temporal resolution. The proposed approach is then extended to a novel problem: including investment in renewable generation as a decision variable, so that planners can make the optimal selection of both conventional and renewable generation. Results are shown for both wind and solar generation. A novel load shaping technique is developed to use demand response to contribute to countering the effect of renewable variability, and its impact on the displacement of conventional generation and on renewable energy integration is illustrated. Finally, the GEP with renewable investment is extended to a 25 year planning horizon which includes the retiring of existing units at the ends of their lifetimes and bringing new generation online in the years it is needed, while still ensuring sufficient intra-hour (5 minute) flexibility. Its performance is demonstrated on the IEEE RTS test case"--