A Study on the Deployment of Mesoscale Chemical Hazard Area Monitoring Points by Combining Weighting and Fireworks Algorithms

Yimeng Shi, Hongyuan Zhang, Zheng Chen, Yueyue Sun, Xuecheng Liu, Jin GuSustainability 2023 · DOI

TL;DR

This paper tackles the challenge of efficiently deploying monitoring points for mesoscale chemical hazards by proposing a novel method that combines a weighted importance model with the fireworks algorithm. This approach aims to optimize coverage, reduce resource waste, and improve early warning capabilities for critical areas.

What Problem Does This Paper Address?

The increasing prevalence of hazardous chemicals, whether from industrial accidents, deliberate acts of terrorism, or geopolitical conflicts, poses a significant threat. Effectively monitoring these threats, especially over mesoscale areas (tens of kilometers), is crucial for public safety. Traditional methods often suffer from resource redundancy and inefficient deployment, leading to wasted resources and potential blind spots. Existing research has largely focused on micro and small-scale hazard monitoring, leaving a gap in optimizing deployment strategies for larger, mesoscale regions where the consequences of delayed detection can be severe. This paper specifically addresses the problem of redundant monitoring point placement and resource waste in mesoscale chemical hazard areas.

Methods

The authors propose a multi-stage approach:

  1. Grid Division and Importance Calculation: The mesoscale monitoring area is divided into a grid. Each grid cell is assigned an "importance" weight based on a quantitative model of chemical hazard ranking.
  2. K-Means Clustering: The gridded areas are then clustered using the k-means algorithm to group cells of similar importance, forming regions with varying levels of criticality. This allows for differentiated monitoring strategies.
  3. Weighted Coverage Model: A mathematical model is developed that considers the importance of each region and aims to achieve a target coverage threshold for each cluster. The model seeks to minimize the number of monitoring nodes (K) and coverage redundancy (Prec).
  4. Fireworks Algorithm: The complex optimization problem of finding the optimal deployment locations is solved using the Fireworks Algorithm. This heuristic algorithm is chosen for its ability to perform global and local searches efficiently, simulating fireworks explosions to explore the solution space.
  5. Sensing Model: A probabilistic sensing model (exponential decay) is used to represent the detection capabilities of the telemetry sensors.

Datasets

Not discussed in the paper. The experiments were conducted using simulated data based on a defined experimental scene (50x50 km chemical hazard area).

Key Insights and Ideas

  • Prioritizing Importance: Not all areas are equally critical. By assigning weights and clustering based on importance, monitoring resources can be strategically allocated to high-risk zones, maximizing their impact.
  • Fireworks Algorithm Efficacy: The fireworks algorithm proves to be a capable tool for solving the complex sensor deployment optimization problem, demonstrating rapid convergence and effective exploration of the solution space.
  • Grid Size Matters: The choice of grid edge length significantly impacts the algorithm's performance, affecting execution time, the number of monitoring points required, and redundancy. A balance needs to be struck for optimal results.
  • Clustering Trade-offs: While the number of clusters has a relatively minor impact on the overall deployment scheme, it influences the granularity of differentiated monitoring strategies.

Key Results

  • The proposed weighted coverage algorithm significantly reduces the number of monitoring points (by 80.5% on average) and redundancy (by 87.5%) compared to an unweighted, indifferent coverage approach.
  • The grid edge length has a substantial influence. For instance, a grid edge length of 3 km resulted in the minimum number of monitoring points (92.6% less than 1 km edge length), 95.7% less redundancy, and a 98.5% reduction in running time, indicating an optimal comprehensive monitoring effect.
  • Increasing the average importance of monitoring areas leads to a higher number of monitoring points and increased redundancy, highlighting the direct correlation between risk and resource allocation.
  • The fireworks algorithm demonstrated its ability to find optimal solutions, reducing redundancy and the number of monitoring nodes while maintaining effective coverage.

Limitations and Future Work

  • Algorithm Choice: The paper primarily relies on the fireworks algorithm. Future work could explore other heuristic algorithms (e.g., simulated annealing, ant colony, genetic algorithms) to find potentially better-suited or more robust solutions.
  • Dimensionality: The current model focuses on a 2D planar area. Extending the research to include a third dimension (height) for spatial monitoring point deployment is a promising future direction.
  • Real-world Data: The experiments were conducted using simulated data. Validation with real-world chemical hazard incident data would further strengthen the findings.
  • Sensor Heterogeneity: The current model assumes homogeneous sensors. Future work could explore deployment strategies for heterogeneous sensor networks with varying capabilities.

Summary updated

Abstract

In order to address the problems of redundancy and waste of resources in the deployment of monitoring points in mesoscale chemical hazard areas, we propose a method for the deployment of monitoring points in mesoscale chemical hazard areas by combining weight and fireworks algorithms. Taking the mesoscale chemical hazard monitoring area as the research background, we take the probabilistic sensing model of telemetry sensor nodes as the research object, make a reasonable grid division of the mesoscale monitoring area, calculate the importance of each grid and perform clustering, utilize the diversity of the fireworks algorithm and the rapidity of the solution to solve the monitoring point deployment model and discuss the relevant factors affecting the deployment scheme. The simulation results show that the proposed algorithm can achieve the optimal coverage monitoring for monitoring areas with different importance and reduce the number of monitoring nodes and redundancy; meanwhile, the relevant factors such as the grid edge length, the number of clusters, and the average importance of monitoring areas have different degrees of influence on the complexity of the algorithm and the deployment scheme.
BibTeX
@inproceedings{shi2023study,
  author = {Yimeng Shi and Hongyuan Zhang and Zheng Chen and Yueyue Sun and Xuecheng Liu and Jin Gu},
  title = {{A Study on the Deployment of Mesoscale Chemical Hazard Area Monitoring Points by Combining Weighting and Fireworks Algorithms}},
  booktitle = {Sustainability},
  year = {2023},
  doi = {10.3390/su15075779},
}
Figure