In vivo evolutionary computation is a novel knowledge-aided, microrobots-oriented tumor targeting framework, where externally manipulable microrobots are employed to detect the cancer in the human vascular network similar to the procedure of solving an optimization problem by swarm intelligence algorithms. The microrobots play the role of computational agents in the optimization procedure, the vascular network is the search space, and the tumor represents the maximum or minimum to be found by agents. Previous work on this topic provided basic computational models and search strategies, which, however, were solely verified in silico. In this letter, we use Janus microparticles as magnetic microrobots, a two-dimensional microchannel network as the human vasculature, and two representative test functions as the exemplar tumor-triggered biological gradient fields to validate in vitro the orthokinetic gravitational search algorithm proposed. The results herein demonstrate the advantages of the algorithm by presenting experimental observations on the targeting performance.

中文翻译：

---

二维微通道网络中基于体内进化计算的微型机器人

体内进化计算是一种新型的知识辅助的、面向微型机器人的肿瘤靶向框架，其中采用外部可操纵的微型机器人来检测人体血管网络中的癌症，类似于通过群体智能算法解决优化问题的过程。微型机器人在优化过程中扮演计算代理的角色，血管网络是搜索空间，肿瘤代表代理要找到的最大值或最小值。之前关于这个主题的工作提供了基本的计算模型和搜索策略，但是，这些都只在计算机上进行了验证。在这封信中，我们使用 Janus 微粒作为磁性微型机器人，使用二维微通道网络作为人体脉管系统，和两个具有代表性的测试函数作为示例肿瘤触发的生物梯度场，以在体外验证提出的正运动重力搜索算法。本文的结果通过展示对目标性能的实验观察来证明该算法的优势。