Distributed security resource allocation strategy based on deep reinforcement learning

Abstract

To achieve reasonable allocation of network security resources, a distributed security resource allocation strategy based on deep reinforcement learning is raised. Firstly, based on the dynamic allocation characteristics of distributed blockchain, a wireless network model of alliance chain is constructed. Subsequently, deep reinforcement learning technology is introduced to optimize the parameters of the alliance chain wireless network model, resulting in a deep reinforcement learning (DRL) based Security Resource allocation model. The simulation comparison results showed that under 100 simulation tests, the cost for high priority user groups was around 1.8 × 10 4 , while the cost for low priority user groups was around 1.2 × 10 4 . In terms of energy efficiency, the energy efficiency of the strengthened security resource allocation model increased from 1.1 Mbps/W to 3.8 Mbps/W. The system transmission rate of the random allocation model increased from 0.2 Mbps/W to 1.6 Mbps/W. The research outcomes denoted that the practical application of the DRL-based Security Resource allocation model in security resource allocation is feasible and effective.