This presentation concerns the promising new area of flexible polygeneration, a chemical process design concept in which a chemical plant is able to change its product outputs throughout its lifetime in response to changing market conditions, business objectives, or other external factors. In this talk we present a new flexible polygeneration process system that can switch between dimethyl ether (DME) or methanol production, depending on need. Classic flexible polygeneration systems typically utilize separate process trains for each product, in which whole process trains are turned on or off (or up or down) depending on the current product. However, our proposed process combines the two process trains into one, in which most of the process equipment is always used during either mode of production, but with different operating conditions. In this work, we show how this significantly reduces capital expenditure, reduces the plant footprint, and ultimately is more economical than a traditional two-train approach. However, the optimal design problem is complicated because it requires both uncertainty considerations and a sufficiently high level of model rigour for the process equipment since the equipment will be utilized in different ways depending on the mode of operation. Therefore, we also present a novel optimization framework and accompanying methodology which can be used to solve the optimal design problem to global optimality without any loss of model rigour (in our case, rigorous Aspen Plus simulations with embedded Aspen Capital Cost Estimator economic predictions). This is achieved through a tabulation approach that exploits process structure for multicomponent distillation. The methodology allows for the rapid solution of many different kinds of optimization formulations, such as robust min-max formulations, scenario-based approaches, and other formulations, based on the amount of predictive knowledge about the future operations of the flexible polygeneration facility known at the conceptual process design stage.