TY - GEN
T1 - A Forward-wave Neural Network for Solving the Priority Shortest Path Problem
AU - Zhu, Haoyu
AU - Lin, Sheng
AU - Zhang, Honghao
AU - Huang, Wei
N1 - Publisher Copyright:
© 2021 ACM.
PY - 2021/10/22
Y1 - 2021/10/22
N2 - In this study, we proposed a forward-wave neural network (FNN) framework that can find the optimal solution to the priority shortest path problem, which is difficult to achieve with traditional algorithms (such as Dijkstra). The basic idea of FNN comes from the following mechanism: the priority shortest path depends on the rules for divide priority and the automatic waves that can reach the destination node. In the design of FNN, each node in the network is regarded as a wave-based neuron. A neuron is composed of six parts, an input, a wave receiver, a wave filter, a wave generator, a wave sender, and an output. Unlike traditional neural networks, FNN does not require any training. The performance evaluation of FNN and Dijkstra algorithm based on the New York Road example proves that FNN is better than Dijkstra algorithm.
AB - In this study, we proposed a forward-wave neural network (FNN) framework that can find the optimal solution to the priority shortest path problem, which is difficult to achieve with traditional algorithms (such as Dijkstra). The basic idea of FNN comes from the following mechanism: the priority shortest path depends on the rules for divide priority and the automatic waves that can reach the destination node. In the design of FNN, each node in the network is regarded as a wave-based neuron. A neuron is composed of six parts, an input, a wave receiver, a wave filter, a wave generator, a wave sender, and an output. Unlike traditional neural networks, FNN does not require any training. The performance evaluation of FNN and Dijkstra algorithm based on the New York Road example proves that FNN is better than Dijkstra algorithm.
KW - Forward-wave neural network
KW - automatic wave
KW - divide priority
KW - the priority shortest path
UR - http://www.scopus.com/inward/record.url?scp=85122697971&partnerID=8YFLogxK
U2 - 10.1145/3501409.3501567
DO - 10.1145/3501409.3501567
M3 - Conference contribution
AN - SCOPUS:85122697971
T3 - ACM International Conference Proceeding Series
SP - 877
EP - 881
BT - Proceedings of the 2021 5th International Conference on Electronic Information Technology and Computer Engineering, EITCE 2021
PB - Association for Computing Machinery
T2 - 5th International Conference on Electronic Information Technology and Computer Engineering, EITCE 2021
Y2 - 22 October 2021 through 24 October 2021
ER -