摘要：This paper proposes novel multiple-mobile-robot collision avoidance path planning based on cooperative co-evolution,which can be executed fully distributed and in parallel. A real valued co-evolutionary algorithm is developed to coordinate the movement of multiple robots in 2D world, avoiding C-space or grid net searching. The collision avoidance is achieved by cooperatively co-evolving segments of paths and the time interval to pass them. Methods for constraint handling, which are developed for evolutionary algorithm, make the path planning easier. The effectiveness of the algorithm is demonstrated on a number of 2Dpath planning problems.

摘要：path planning in 3D geometry space is used to find an optimal path in the restricted environment, according to a certain evaluation criteria. To solve the problem of long searching time and slow solving speed in 3D path planning, a modified ant colony optimization is proposed in this paper. Firstly, the grid method for environment modeling is adopted. Heuristic information is connected with the planning space. A semi-iterative global pheromone update mechanism is proposed. Secondly, the optimal ants mutate the paths to improve the diversity of the algorithm after a defined iterative number. Thirdly, co-evolutionary algorithm is used. Finally, the simulation result shows the effectiveness of the proposed algorithm in solving the problem of 3D pipe path planning.

摘要：Space syntax has proven there appears to be a fundamental process that informs human and social usage of an environ- ment, and the effects of spatial configuration on movement patterns are consistent with a model of individual decision behavior. In- troducing space syntax to crowd simulation will enable space structure guide the random movement of the crowd with no specific targets. This paper proposes a simulation method of the wandering crowd, which calculates the crowd distribution corresponding to the space through space syntax and uses a hybrid path planning algorithm to dynamically navigate the crowd to conform to the dis- tribution. Experiments show the presented method can obtain reasonable and vision-realistic simulation results.