A parallel time‐varying earliest arrival path algorithm for evacuation planning of underground mine water inrush accidents

Abstract

Evacuation planning is of importance to reduce casualties in underground mine water inrush accidents. Since the transit time depends on the water height along the spreading of the water inrush, finding a path with the minimum egress time through the roadway network is considered as a time-varying shortest path problem. Such problem can be solved by the time-varying earliest arrival path algorithm. However, although the algorithm can solve the problem in polynomial time, more improvements are demanded to meet the real-time requirements of practical applications.In this study, a parallel time-varying earliest arrival path algorithm is proposed in this regard. The vertex-hyperedge network is recommended to represent the adjacency of vertices. Three parallel strategies, including the time-based decomposition, adjacent-vertex-based decomposition, and the adjacent-vertex-based decomposition using vertex-hyperedge network, are proposed for parallelism implementation. The OpenMP interface is employed to evaluate the performance of the proposed strategies on a share-memory multi-core computer. The experiment resultsshow that all three strategies can decrease the run time of the algorithm while maintaining the reasonable accuracy. The best strategy is adjacent-vertex-based decomposition using vertex-hyperedge network, which results in the 2-fold speedup when the number of threads meets the degree of concurrency of the topology structure. KEYWORDS evacuation planning, mine water inrush accidents, parallel earliest arrival path algorithm, vertex-hyperedge network INTRODUCTIONDetermining the escape route during the evacuation plays an important role in decreasing casualties in case of water inrush accidents in underground mines. 1,2 Considering the space limitations in underground tunnels, roadways are submerged along the spreading progress of the mine water inrush. Water inrush to the mining tunnel caused many tragic moments so far. The most catastrophic accident in the history happened when the whole roadways was submerged. In these accidents, when miners simply escape along the shortest or empirical route to the exit, they may face the dead end path and get drown. Researchers applied the computational fluid dynamics in the geometrical method and proposed few water inrush spreading models to simulate the water inrush in the underground tunnel and predict the extend of the water filling in roadways. 3These models can be applied to compute the water height of roadways over time. The traverse time of each roadway can be predicted by empirical functions for the wading speed and water height. Therefore, when the traverse time is considered as the edge weight, roadways of the whole underground mine change from a static network to a time-varying network. It should be indicated that such a roadway network is usually a sparse network, because roadways are exploited along up and down coal seams. Therefore, traverse points are placed into the roadway network Concurrency Computat Pract Exper. 2020;32:e5644. wileyonlinelibrary.com/journ...

A genetic algorithm‐based intelligent solution for water pipeline monitoring system in a transient state

“…Consequently, it helps reduce water wastage and gives one step ahead of smart cities. 6,7 Accordingly, the Structural Health Monitoring of pipelines 8 is important for maintaining an accurate water pipeline system (WPS) 9 and detecting and localizing any possible hydraulic damage.…”

A cloud‐based service for emergency evacuation of mine water inrush accidents

“…These four methods can solves main problem of capacity‐based path planning in the view of macroscopic planning. In addition, since edge weight can varied over time, Wu et al 7,8 propose a time‐varying shortest path algorithm to find the earliest arrival path and its parallel version for searching the shortest route through a network with edge weight variation over time with a better computation efficiency.On the other hand for the perspective of microscopic planning, agent‐based methods are adopted to simulate crowd behavior. Pan et al 9 present an agent‐based framework to simulate human and social behaviors during emergency evacuations.…”

“…These four methods can solves main problem of capacity‐based path planning in the view of macroscopic planning. In addition, since edge weight can varied over time, Wu et al 7,8 propose a time‐varying shortest path algorithm to find the earliest arrival path and its parallel version for searching the shortest route through a network with edge weight variation over time with a better computation efficiency.…”