Decision making under uncertainty is becoming more important in process industries as optimisation is applied to novel applications as well as plant-wide and enterprise optimisation. Among the standard stochastic optimisation techniques are stochastic programming and dynamic programming. It is difficult to use these techniques for practical applications due to unwieldy computational requirements, arising from a large number of uncertain parameters and state variables, respectively. In this paper, we present sample-based techniques for ameliorating the computational difficulties. Application studies involving catalyst design and real-time optimisation point to the promising potentials of the sample-based techniques.