MSGW-FLM] Optimizing Emergency Rescue In Disasters With The Power Of IoT

Internet Of Things 17/06/2024

3 main points
✔️ This model optimizes emergency rescue resource allocation, taking into account a variety of constraints and objectives.
✔️ As a result, it is noteworthy that MSGW-FLM performs better than other algorithms, particularly in achieving system losses that ultimately approach zero. These findings highlight the effectiveness and suitability of MSGW-FLM in managing emergency material assignment tasks.
✔️ Future work will validate the effectiveness of this model in real-world emergency scenarios and further test its ability to dynamically update spatio-temporal information.

A Multi-constraint and Multi-objective Allocation Model for Emergency Rescue in IoT Environment
written by Xinrun Xu, Zhanbiao Lian, Yurong Wu, Manying Lv, Zhiming Ding, Jian Yan, Shang Jiang
(Submitted on 15 Mar 2024)
Comments: 5 pages, 5 figures, ISCAS 2024
Subjects: Artificial Intelligence (cs.AI)

code：

The images used in this article are from the paper, the introductory slides, or were created based on them.

Summary

Effective resource allocation is necessary in the immediate aftermath of a disaster. Timely, informed decision making is critical, especially in long-term crises and large-scale disasters. Recently, leveraging advances in IoT and spatio-temporal data analysis, a system called the Multi-Objective Shuffle Gray Wolf Frog Leap Model (MSGW-FLM) has been developed. This model optimizes emergency rescue resource allocation, taking into account a variety of constraints and objectives. Tests in real emergency rescue scenarios have shown better performance than existing models, and MSGW-FLM represents a significant advance in optimizing resource allocation in complex situations.

Introduction

Rapid advances in artificial intelligence, IoT sensing, and smart cities have focused attention on the integration of these technologies for post-disaster emergency response. Emergency response is critical to protect life and property, and complex information must be analyzed to develop effective rescue strategies. Spatio-temporal data is essential for this purpose, and IoT devices enable real-time data collection and the proper allocation of emergency resources. Furthermore, rescue plans can be modified to adapt to changing conditions for more effective rescue operations. Such data-driven emergency response is a dynamic and iterative process in which real-time data from IoT devices plays a critical role. In particular, the use of IoT with spatio-temporal data is of great importance in post-disaster emergency response, and new approaches based on this have been proposed.

MSGW-FLM

In this paper, a novel approach combining multiple meta-heuristic algorithms is proposed as a solution to the constrained multi-objective decision optimization problem (CMDOP). Specifically, the gray wolf optimization algorithm (GWOA) and the shuffled frog leap algorithm (SFLA) are combined to develop a multi-objective shuffled gray wolf-frog leap model (MSGW-FLM). The following figure is a schematic diagram of the system.

SFLA is a meta-heuristic approach based on frog foraging behavior, where frogs are grouped into memeplexes to exchange information and reinforce solutions. GWOA, on the other hand, is an iterative optimization algorithm inspired by the rank and hunting scenario of the gray wolf. The combination of these algorithms allows for efficient solution space exploration.

Combined, these algorithms allow MSGW-FLM to efficiently search the solution space and improve the accuracy of solution ranking in multi-objective problems. Furthermore, the method has been demonstrated to perform better than existing optimization methods such as NSGA-II and congestion distance.

MSGW-FLM is also flexible enough to be applied to different problem domains and to complex decision-making problems that consider multiple constraints and objectives. This enables more efficient and appropriate search for solutions in a variety of real-world problems.

Experiment

We will evaluate the proposed method, MSGW-FLM. First, we design two material allocation examples with multiple objectives and use MSGW-FLM to solve the associated decision models. Specifically, we will assume a dynamic decision-making scenario and employ a rolling time-domain planning approach. With this approach, the decision is divided into multiple cycles, with MSGW-FLM being utilized in each cycle to execute the plan. As new data emerges, it is incorporated into the next cycle and model parameters are dynamically adjusted.

We then evaluate MSGW-FLM against other algorithms. To do this, we use four classical benchmark problems and evaluate 28 multiobjective subproblems. Each algorithm is evaluated using performance metrics such as HV, IGD, and Spread, which show that MSGW-FLM outperforms the other baseline algorithms in many cases. The following table shows the HV, IGD, and Spread performance of the optimized algorithms in ZDT, WFG, DTLZ, and LZ09 F.

To further evaluate the capabilities of MSGW-FLM, we delve into scenarios with unpredictable demand and supply dynamics. Specifically, we consider the allocation of materials between distribution centers and disaster areas with random supply and demand points. In this setting, there are 10 designated disaster areas and 10 distribution centers over five planning cycles. Each cycle determines the ambiguous demand for all disaster sites and the uncertain supply for each distribution center at each planning stage. This allows us to visually investigate the average change in total system losses for different supply/demand point combinations. The following graphs show the actual results presented within the paper.

As a result, it is noteworthy that MSGW-FLM performs better than other algorithms, especially in achieving system losses that ultimately approach zero. These findings highlight the effectiveness and suitability of MSGW-FLM in managing emergency material assignment tasks.

Conclusion

This study investigated the use of spatio-temporal data in the IoT to account for multiple objectives and constraints in the effective allocation of emergency resources. For this purpose, a pioneering approach, the Multi-Objective Shuffling Shuffle Gray Wolf Frog Leaping Model (MSGW-FLM), was introduced. The model was then tested on 28 multiobjective tasks using spatio-temporal data. Results showed that MSGW-FLM consistently outperformed other baseline models such as NSGA-II, IBEA, and MOEA/D. Future studies will validate the effectiveness of this model in real-world emergency scenarios and further test its ability to dynamically update spatio-temporal information.