An evolutionary multi-objective path planning of a fleet of ASVs for patrolling water resources
The rapid increase of human activities with direct influence on the environment has motivated the global awareness of the need to efficiently monitor the natural resources. Among the wide range of problems addressed, such as overuse of agrochemicals, uncontrolled waste, etc., the contamination of wa...
| Autores: | , , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2022 |
| País: | España |
| Institución: | Universidad de Sevilla (US) |
| Repositorio: | idUS. Depósito de Investigación de la Universidad de Sevilla |
| OAI Identifier: | oai:idus.us.es:11441/135124 |
| Acceso en línea: | https://hdl.handle.net/11441/135124 https://doi.org/10.1016/j.engappai.2022.104852 |
| Access Level: | acceso abierto |
| Palabra clave: | Autonomous surface vehicles Water monitoring Genetic algorithm Patrolling problem Multi-objective optimization |
| Sumario: | The rapid increase of human activities with direct influence on the environment has motivated the global awareness of the need to efficiently monitor the natural resources. Among the wide range of problems addressed, such as overuse of agrochemicals, uncontrolled waste, etc., the contamination of water resources plays a protagonist role, given its close links with biodiversity and the food chain. Water monitoring is considered one of the most efficient ways to deal with these problems, especially through the use of autonomous vehicles, which can boost the capabilities and efficiency of the monitoring routines with appropriate strategies. In this work, the monitoring problem is addressed by means of the Non-Homogeneous Patrolling Problem with closed circuits. This problem has a great computational complexity, especially when multiple targets are included in a monitoring mission. A formulation based on closed metric graphs and the application of a multi-objective genetic algorithm is proposed to provide Pareto-efficient monitoring solutions for a variable number of Autonomous Surface Vehicles. To address the multi-agent, multi-objective and constrained paradigm, efficient genetic operators have been designed for the generation of valid solutions in an affordable time. The method results in Pareto-efficient solutions for scenarios with disjoint and uncorrelated objectives, which outperform the fitness of other solutions by a factor of 2, on average. The results provide decision makers a method to find different non-dominated strategies depending on the monitoring needs, depending on fleet and vehicle size. |
|---|