Background and Objective: Cancer tumors constitute a complicated environment for conventional anti-cancer treatments to confront, so solutions with higher complexity and, thus, robustness to diverse conditions are required. Alternations in the tumor composition have been documented, as a result of a conventional treatment, making an ensemble of cells drug resistant. Consequently, a possible answer to this problem could be the delivery of the pharmaceutic compound with the assistance of nano-particles (NPs) that modify the delivery characteristics and biodistribution of the therapy. Nonetheless, to tackle the dynamic response of the tumor, a variety of application times of different types of NPs could be a way forward. Methods: The in silico optimization was investigated here, in terms of the design parameters of multiple NPs and their application times. The optimization methodology used an open-source simulator to provide the fitness of each possible treatment. Because the number of different NPs that will achieve the best performance is not known a priori, the evolutionary algorithm utilizes a variable length genome approach, namely a metameric representation and accordingly modified operators. Results: The results highlight the fact that different application times have a significant effect on the robustness of a treatment. Whereas, applying all NPs at earlier time slots and without the ordered sequence unveiled by the optimization process, proved to be less effective. Conclusions: The design and development of a dynamic tool that will navigate through the large search space of possible combinations can provide efficient solutions that prove to be beyond human intuition.

背景与目的：癌症肿瘤构成了常规抗癌治疗面临的复杂环境，因此需要具有更高复杂性的解决方案，因此需要对各种状况的鲁棒性。作为常规治疗的结果，已经证明了肿瘤组成的变化，使细胞整体具有抗药性。因此，该问题的可能答案可能是借助纳米颗粒（NPs）来输送药物化合物，从而改变治疗的输送特性和生物分布。尽管如此，为了应对肿瘤的动态反应，不同类型NP的多种应用时间可能是前进的方向。方法：计算机模拟根据多个NP的设计参数及其应用时间，对优化进行了研究。优化方法使用开源模拟器来提供每种可能治疗方法的适用性。由于先验未知将获得最佳性能的不同NP的数量，因此进化算法利用了可变长度基因组方法，即同分异构表示法和相应的修饰算子。结果：结果突出了这样一个事实，即不同的应用时间对治疗的坚固性具有重要影响。然而，在较早的时隙应用所有NP且未通过优化过程显示有序序列的事实证明效果不佳。结论： 动态工具的设计和开发将在可能的组合的大型搜索空间中导航，可以提供证明超出人类直觉的有效解决方案。