Organic structure-directing agents (OSDAs) are often employed for synthesis of zeolites with desired frameworks. A priori prediction of such OSDAs has mainly relied on the interaction energies between OSDAs and zeolite frameworks, without cost considerations. For practical purposes, the cost of OSDAs becomes a critical issue. Therefore, the development of a computational de novo prediction methodology that can speed up the trial-and-error cycle in search for less expensive OSDAs is desired. This study utilized a nature-inspired ant colony optimization method to predict physicochemically and/or economically preferable OSDAs, while also taking molecular similarity and heuristics of zeolite synthesis into consideration. The prediction results included experimentally known OSDAs, candidates having structures closely related to known OSDAs, and novel ones, suggesting the applicability of this approach.

有机结构导向剂（OSDA）通常用于合成具有所需骨架的沸石。对此类OSDA的先验预测主要依赖于OSDA与沸石骨架之间的相互作用能，而没有考虑成本。出于实际目的，OSDA的成本成为一个关键问题。因此，开发了从头计算需要一种可以加快试验和错误周期以寻找较便宜的OSDA的预测方法。这项研究利用自然界启发性的蚁群优化方法来预测理化和/或经济上优选的OSDA，同时还考虑了分子相似性和沸石合成的启发法。预测结果包括实验已知的OSDA，具有与已知OSDA密切相关的结构的候选对象以及新颖的候选对象，这表明该方法的适用性。