The continuous expansion of smart grid substations has led to increasingly complex communication network topologies

with the number of nodes and links growing exponentially and their interconnections significantly strengthened. The state of links is influenced by various factors such as equipment failures

environmental interference

and load fluctuations

exhibiting highly dynamic characteristics. Traditional static path selection methods rely solely on simple rule-based judgments for path selection

lacking precise awareness of the global network state and struggling to adapt in real-time to the dynamic changes in link states. This results in an inability to efficiently and accurately identify the optimal communication path

severely constraining the reliable operation of the smart grid. To address these issues

this paper proposes an ultra-low latency communication path optimization scheme based on an improved ant colony algorithm. By thoroughly analyzing the characteristics of substation communication networks

constraints on data transmission delay and link operational states are established

with ultra-low latency as the optimization objective. Based on graph theory

a substation ultra-low latency communication path model is constructed

using adjacency matrices and weight matrices to provide an accurate mathematical description of the complex network topology

laying the foundation for subsequent path optimization. A dynamic pseudo-random transfer rule is introduced to overcome the limitations of traditional fixed rules

enhancing adaptability of the algorithm to the uncertainty of data transmission. Additionally

a dual-layer pheromone update mechanism is designed: local pheromone updates employ a dynamic decay factor to reflect real-time changes in link states

while global pheromone updates incorporate an elite strategy to accelerate the convergence speed of high-quality paths. Through multiple rounds of iterative searches

the path selection strategy is continuously optimized

ultimately yielding an ultra-low latency communication path. Experimental results show that

compared to traditional methods

the proposed scheme achieves the shortest communication data transmission path and the smallest corresponding communication delay

with throughput consistently maintained above 65 MB

significantly enhancing network throughput. Furthermore

when the number of communication nodes is 130

the network connectivity remains above 0.55

effectively reducing unnecessary bandwidth usage.