The economic and efficient scheduling of electrical generators is a complex optimization problem. The problem becomes even more complex when the uncertain solar sources are considered. Over the last decade, different scenario-based solution approaches have been developed for this problem in which the uncertain solar productions are represented as random scenarios. Although this approach is beneficial for a day-ahead scheduling, the selection of an appropriate number of scenarios is very challenging. A large number of scenarios is required to obtain a stable solution, but it takes longer computational time to solve. In this paper, a dynamic scenariobased optimization model is developed in which the number of scenarios is dynamically set during the process of evaluations. The optimization model is solved using a differential evolution in which the control parameters are self-adaptively adjusted for better performance. A 19-unit solar thermal problem for a 24-hour time horizon is taken from the literature and solved using the proposed and conventional approaches. The obtained results are analyzed that shows that the proposed approach has some merits in terms of computational efficiency and solution stability.