The ongoing deregulation of electricity industries worldwide is providing incentives for microgrids, entities that use small-scale distributed generation (DG) and combined heat and power (CHP) ap- plications to meet local energy loads, to evolve independently of the traditional centralised grid in order to provide greater flexibility and energy efficiency to end-use consumers. We examine the impact of start-up costs on the option values and operating schedules of on-site DG installed by a microgrid in the presence of stochastic electricity and fuel prices. We proceed by formulating a stochastic dynamic programme (SDP) for the microgrid that minimises its expected discounted cost over a time horizon and solving it using least-squares Monte Carlo (LSMC) simulation. The expected cost saving that the microgrid realises by having gas-fired DG installed relative to meeting its entire electric load via off-site purchases is the implied option value of DG. Numerical examples indicate that although start-up costs do not significantly lower DG value, they, nevertheless, have a profound impact on the optimal DG operating schedule as the microgrid must incorporate not only current, but also future, expected start-up costs into its current decision-making process as an opportunity cost. As a consequence, the microgrid becomes more hesitant to turn DG units on (off), preferring to wait until the electricity price (natural gas generating cost) exceeds the natural gas generating cost (electricity price) by a significant margin before taking action. We demonstrate that ignoring this tradeoff and proceeding myopically as in the case without start-up costs results in drastically higher expected costs and fewer opportunities to use DG.