Abstract
Many transport systems in the real world can be modeled as networked systems. Due to limited resources, only a few nodes can be selected as seeds in the system, whose role is to spread required information or control signals as widely as possible. This problem can be modeled as the influence maximization problem. Most of the existing selection strategies are based on the invariable network structure and have not touched upon the condition that the network is under structural failures. Related studies indicate that such strategies may not completely tackle complicated diffusion tasks in reality, and the robustness of the information diffusion process against perturbances is significant. To give a numerical performance criterion of seeds under structural failure, a measure has been developed to define the robust influence maximization (RIM) problem. Further, a memetic optimization algorithm (MA) which includes several problem-orientated operators to improve the search ability, termed , has been presented to deal with the RIM problem. Experimental results on synthetic networks and real-world networks validate the effectiveness of , its superiority over existing approaches is also shown.
Highlights
Automatic guided vehicles (AGVs), which belong to the category of wheeled mobile robots, play a significant role in transportation, the logistics industry, and autonomous driving [2], and can be modeled as networked systems
How to use the topological information of a specific network to select the seeds that can achieve the optimal propagation effect is defined as the influence maximization problem [5], which is of great significance in both theoretical and realistic applications
It can be seen that in three artificial synthesis networks with different scales, the seed set selected by all evolutionary algorithms including RIMMA, memetic optimization algorithm (MA)-sim, and genetic algorithm (GA) achieve a higher Rs value than the seed set obtained by other algorithms
Summary
The Monte Carlo process is optional to evaluate the influence performance σ (S) of the seed set S [8], but this method is time-consuming and may not get accurate estimation results. This simulation is carried out 1000 times, sum and average all σn (S) (divided by n = 1000), and the calculated average is the influence performance of the seeds This method can only deal with evaluation tasks on small-scale networks. For improving the efficiency of the performance estimation process, Lee et al in [11] proposed a fast approximation method for influence spreading; only the influence within the 2-hop range of seeds is considered, defined as follows:. S∈S s∈S c∈Cs ∩S where Cs is the 1-hop neighbor of node s, p(s, c) is the propagation probability from active node s to inactive node c, χ represents the overlapped influence when the influence is estimated between two seeds. In the IC model, node 9 has a fixed probability p to activate node 1
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
