Features of using swarm intelligence algorithms for drone route optimization
Uloženo v:
| Název: | Features of using swarm intelligence algorithms for drone route optimization |
|---|---|
| Zdroj: | Науковий вісник НЛТУ України, Vol 35, Iss 2 (2025) |
| Informace o vydavateli: | Ukrainian National Forestry University, 2025. |
| Rok vydání: | 2025 |
| Témata: | Particle Swarm Optimisation (PSO), Cuckoo Search Algorithm (CSA), Forestry, SD1-669.5, Sparrow Search Algorithm (SSA), Artificial Bee Colony (ABC) |
| Popis: | This paper investigates the use of swarm algorithms for optimising delivery routes using Unmanned Vehicles (UVs) as an effective alternative to traditional algorithms. Classical path planning algorithms such as Dijkstra's algorithm, A* or APF (Artificial Potential Field) can struggle in real-world environments due to the ability to fall into local minima, or due to increased computational complexity in cases with multiple obstacles. To avoid these limitations, the research focuses on path-planning methods through swarm algorithms, many of which are inspired by animal biological behaviour. The article considers the features of operation and suitability for use in dynamic conditions of such algorithms as Particle Swarm Optimisation (PSO), Artificial Bee Colony (ABC), Cuckoo Search Algorithm (CSA) and Sparrow Search Algorithm (SSA). The study showed that there is no one universal algorithm for all types of environments. All algorithms demonstrate high performance when used under the best conditions. An algorithm's effectiveness depends on considering weaknesses and strengths when planning a task. It was found that PSO shows fast convergence even when using a large number of agents, which makes it suitable for route planning in swarm systems. Instead, the extensibility of the ABC algorithm and its ease of computational complexity allows for the efficient distribution of tasks among agents in large swarm systems. The CSA allows for the best local planning, which, in combination with other algorithms, allows for the best results in the balance between global and local route planning. SSA achieves fast convergence at the cost of computational resources, while ensuring execution in a dynamic environment. The results of the study show that the best algorithms for dynamic environments are those that combine local and global search capabilities, such as PSO modifications. The study emphasises the role of swarm intelligence algorithms in improving UAVs' performance by providing robust, efficient and scalable delivery routes. Future research could concentrate on exploring new hybrid algorithm models that combine swarm intelligence algorithms with machine learning technologies to improve algorithm outcomes in dynamic environments. In order to enhance the algorithms and analyse their limitations, further testing of the algorithms in complex environments with a large number of interferences is important to make further improvements to the algorithms. |
| Druh dokumentu: | Article |
| ISSN: | 2519-2477 1994-7836 |
| DOI: | 10.36930/40350212 |
| Přístupová URL adresa: | https://doaj.org/article/df3338a2a01445909799d6d21c30dc9e |
| Rights: | CC BY |
| Přístupové číslo: | edsair.doi.dedup.....0e264956d6225650e62adb86a358fac5 |
| Databáze: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstrakt: | This paper investigates the use of swarm algorithms for optimising delivery routes using Unmanned Vehicles (UVs) as an effective alternative to traditional algorithms. Classical path planning algorithms such as Dijkstra's algorithm, A* or APF (Artificial Potential Field) can struggle in real-world environments due to the ability to fall into local minima, or due to increased computational complexity in cases with multiple obstacles. To avoid these limitations, the research focuses on path-planning methods through swarm algorithms, many of which are inspired by animal biological behaviour. The article considers the features of operation and suitability for use in dynamic conditions of such algorithms as Particle Swarm Optimisation (PSO), Artificial Bee Colony (ABC), Cuckoo Search Algorithm (CSA) and Sparrow Search Algorithm (SSA). The study showed that there is no one universal algorithm for all types of environments. All algorithms demonstrate high performance when used under the best conditions. An algorithm's effectiveness depends on considering weaknesses and strengths when planning a task. It was found that PSO shows fast convergence even when using a large number of agents, which makes it suitable for route planning in swarm systems. Instead, the extensibility of the ABC algorithm and its ease of computational complexity allows for the efficient distribution of tasks among agents in large swarm systems. The CSA allows for the best local planning, which, in combination with other algorithms, allows for the best results in the balance between global and local route planning. SSA achieves fast convergence at the cost of computational resources, while ensuring execution in a dynamic environment. The results of the study show that the best algorithms for dynamic environments are those that combine local and global search capabilities, such as PSO modifications. The study emphasises the role of swarm intelligence algorithms in improving UAVs' performance by providing robust, efficient and scalable delivery routes. Future research could concentrate on exploring new hybrid algorithm models that combine swarm intelligence algorithms with machine learning technologies to improve algorithm outcomes in dynamic environments. In order to enhance the algorithms and analyse their limitations, further testing of the algorithms in complex environments with a large number of interferences is important to make further improvements to the algorithms. |
|---|---|
| ISSN: | 25192477 19947836 |
| DOI: | 10.36930/40350212 |