This thesis studies the maintenance and replacement decision problem for systems subject to uncertain technological improvement while considering the operational requirements related to maintenance and strategic investment. We utilize the flexibility and dynamic nature of real options that traditionally applied to the financial domain by evaluating them through non-stationary Markov decision processes. More specifically, the models presented allow us to dynamically adapt the maintenance decision for systems subject to substantial technological developments. This decision is linked to the notion of a technological jump, the choice of technology generation for investment, as well as the uncertainty of the economy and market. The maintenance option allows delaying the investment decision in order to wait for changes in technology. Furthermore, we analyze these decisions based on different kinds of the obsolescence. They can be subjective, defined by the ratio of current and expected performance, or objective in the case of the material incompatibility, in particular spare parts. Finally, the use of these models allows us to study the influence of the quality and level of technology information on maintenance and investment policies by evaluating the possibility to periodically purchase additional information.
