In the transition toward a circular and CO2-neutral economy, chemical refineries are presented with the urgent need to reinvent themselves. While the demand for fossil-derived fuels will drastically decline in the next decade, the demand for chemical building blocks (light olefins, aromatics,…) is only expected to further increase. In this proposal, the chemistry of two emerging technologies will be investigated, namely (i) direct conversion of crude oil to chemicals and (ii) recycling of plastic waste. Both technologies take place through catalytic cracking on (hierarchical) zeolite catalysts, yet their performance is not yet well understood and reaching economic viability is a major challenge. Their success critically depends on the design of new zeolites to convert these complex feedstocks into the targeted product distribution with high selectivity. The goal of this project is to improve our mechanistic understanding of cracking long-chain hydrocarbons (>C8) and to screen strategies to enhance the light olefin selectivity by modifying the acidity and porosity of the zeolite catalyst. To this end, state-of-the-art molecular simulations, combined with machine-learning techniques, will be performed at actual operating conditions. Due to the long hydrocarbon chains in the feed, new (mesoporous) zeolite models will be built to properly describe all host-guest interactions and innovative diffusion-kinetics models will be established to properly predict the product distribution.