Congestion due to an irregular flow of people has been commonly observed at various spots. Congestion increases unpleasantness and may cause trouble. Moreover, it may lead to a serious accident such as a sudden collapse of the overcrowded structure of people. In order to understand the mechanism of congestion due to the irregular flow of people, we performed a multiagent simulation of the flow, which can be used to analyze various systems of complexity. First, we performed simulations between individuals and simulated the crowd as an application. We simulated lane formation that was a characteristic of the flow of people at a congested state. Then, we simulated the flow of people in the catwalk of a station. It was shown that features of congestion were changed by the positions of obstacles such as columns.