Modified collective decision optimization algorithm with application in trajectory planning of UAV

• Published in: Applied intelligence (Dordrecht, Netherlands) Vol. 48; no. 8; pp. 2328 - 2354
• Main Authors: Zhang, Qingyang, Wang, Ronggui, Yang, Juan, Ding, Kai, Li, Yongfu, Hu, Jiangen
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.08.2018; Springer Nature B.V
• Subjects: Archives & records; Artificial Intelligence; Bonding agents; Computer Science; Computer simulation; Decision making; Evolutionary algorithms; Heuristic methods; Human behavior; Innovations; Machines; Manufacturing; Mathematical analysis; Mechanical Engineering; Operators; Optimization; Optimization algorithms; Processes; Sequences; Trajectory planning; Unmanned aerial vehicles; Path planning; Decision-making behaviour; Unmanned aerial vehicle; Collective decision optimization algorithm
• ISSN: 0924-669X; 1573-7497
• DOI: 10.1007/s10489-017-1082-1

Recently, a new heuristic method called collective decision optimization algorithm (CDOA) was proposed. This paradigm is inspired from the decision-making behaviour of human beings, including the different factors influencing the decisions, such as experience, opinion of others, group thinking, opinion of the leader, and innovation. However, the original version of the algorithm concentrated only on a fixed evolution order. This study introduces an extended version of the CDOA (ECDOA) for the evolution mechanism without making any major conceptual change to its architecture. In ECDOA, all the agents break the bonds of the original operator sequence. With the exception of the innovation operator, all the other operators are initially stored in an external archive, and then each agent combines at least one randomly selected operator from this archive with an innovation operator to create new update orders. This method not only provides more calculation sequences in each iteration, but also generates more promising candidate solutions. In addition, several operators are further modified to improve the optimization abilities. A comprehensive set of modern benchmark functions and UAV path planning are required to verify the effectiveness of the ECDOA thoroughly. The results of the series-simulation comparison, which simultaneously consider both the convergence and accuracy, indicate that ECDOA is more effective and feasible than the other state-of-art optimization paradigms.