A Study of Design and Implementation of Interactive Algorithm

在这个研究项目中，一个新的概念，所谓的“交互式算法”，其中一些上级功能是通过几个系统之间的相互作用，包括人的主张。具体而言，考虑了五种类型的系统，即计算机，人，机器人，优化算法和环境。首先，研究了一种能够区分多种不同环境并完成给定任务的智能机器人。通过与周围环境的交互，机器人不仅能够识别当前环境，而且能够成功地识别自己的位置，而机器人没有地图来表示它将要行动的世界。其次，为了实现人工系统的突现设计方法，开发了人机协同设计工具交互式遗传算法查看器（IGAV），并将其应用于移动的机械手的优化设计。最后，提出了一种将遗传算法的全局搜索能力与传统局部优化算法的局部搜索能力有效结合的元启发式策略。

In this research project, a new concept so-called "interactive algorithm" is advocated, where some superior function is emerged through the interaction among several systems including a human being. Concretely, five types of systems are considered, namely, computer, human, robot, optimization algorithm, and environment. First of all, an intelligent robot that distinguishes several different environments and completes a given task is studied. Through the interaction with surrounding environment, the robot recognizes not only current environment but also its own position successfully, whereas the robot has no map representing the world in which it is going to act. Next, in order to realize the emergent approach for design of artificial systems, a man-machine cooperative design tool, Interactive Genetic Algorithm Viewer (IGAV), is developed and applied to the optimal design of mobile manipulators. Finally, a meta-heuristic strategy that combines the global search power of GA with the local search power of conventional local optimization algorithm effectively is proposed.