Identical components are considered, which become obsolete once new-type ones are available, more reliable and less energy consuming. We envision different possible replacement strategies for the old-type components by the new-type ones: either purely preventive, where all old-type components are replaced as soon as the new-type ones are available; either purely corrective, where the old-type ones are replaced by new-type ones only at failure; or a mixture of both strategies, where the old-type ones are first replaced at failure by new-type ones and next simultaneously preventively replaced after a fixed number of failed old-type components.To evaluate the respective value of each possible strategy, a cost function is considered, which represents the mean total cost on some finite time interval [0, t]. This function takes into account replacement costs, with economical dependence between simultaneous replacements, and also some energy consumption (and/or production) cost, with a constant rate per unit time.A full analytical expression is provided for the cost function induced by each possible replacement strategy. The optimal strategy is derived in long-time run. Numerical experiments conclude the paper. fixed time, say time 0, some new components appear in the market, issued from a new technology, which makes them more reliable, less energy consuming and more performing. Such new-type of components may be substituted to the older ones with no problem of compatibility. There is no stocking of old-type components and after time 0, no old-type component is available anymore (or the industrialist is not allowed to use old-type components anymore, e.g. for safety reasons). After time 0, any failed component, either old-type or new-type, is then instantaneously replaced by a new-type one. At time 0, each old-type component is in use since some random time, with some random remaining lifetime. If the new-type components are much less energy consuming than the older ones and if the period of interest is very long, it may then be expedient to remove all old-type components immediately at time 0 and replace them by new-type ones, leading to some so-called purely preventive replacement strategies. On the contrary, in case there is no much improvement between both technologies and if the period of interest is short, it may be better to wait until the successive failures of the old-type components and replace them by new-type ones only at failure, leading to some purely corrective replacement strategy. More generally, some mixture of both strategies, preventive and corrective, may also be envisioned and may lead to lower costs, as can be seen later. The point of the present paper is to look for the optimal replacement strategy among the purely preventive one, the purely corrective one and the mixtures of both strategies envisioned here. To evaluate the respective value of each possible strategy, a cost function is considered, which represents the mean total cost on some finite time interval [0, t]. This function takes into a...