Many studies have focused on animal behaviors, especially trajectories of their movements such as walking or flying, because the trajectories reflect their decision making based on interactions between internal (information processing) and external (environment) elements. The previous main methodology for analyzing trajectories has been to select the fittest model with stochastic processes (e.g. random walk or Levy walk). However, Maye et al. (2007) reported that spontaneous behaviors exhibited the nonlinear sign rather than stochastic, and the brain which produced the behavior was a nonlinear system. Therefore, the analysis on trajectories as nonlinear dynamics must be useful and lead to the more exact description of the dynamics.

In this study, we investigate the host-searching behavior of the female parasitoid wasp, Heterospilus prosopidis which oviposits on a seed beetle host, Callosobruchus chinensis, using a video tracking system. We propose nonlinear time-series analysis on the trajectories. Additionally, we have developed a virtual agent that has a neural network inside and can interact with cues in its environment. We also analyze the neural dynamics and trajectories of the agents.We discuss how animals interact with environment and why their brain has evolved as nonlinear systems.