Scholarlink

<p>Wireless sensor networks (WSNs) face persistent challenges in energy efficiency and fault tolerance due to limited node battery life. This paper introduces a novel multi-goal intelligent agent approach (iAFTA) for static WSNs, where intelligent agents (iAgents) proactively monitor cluster head (CH) energy, detect failures, and dynamically elect replacements based on energy and proximity. The process integrates fault tolerance to prevent cascading failures while minimizing energy consumption through adaptive minimum spanning tree (MST) reconfiguration. The proposed iAFTA approach is evaluated across five distinct scenarios using key metrics such as first node dead (FND), half node dead (HND), last node dead (LND), average energy consumption (AEC), and CH replacement rate. In addition, fault-tolerance metrics, including fault detection accuracy, detection time, and energy overhead, are evaluated. Results show that iAFTA achieves 100 % fault detection accuracy, immediate detection and recovery (average detection time of 1.0 round), and low-energy overhead during fault handling, even in large-scale scenarios. Results demonstrate significant improvements over existing protocols like BWOA-V. In Scenarios #1 and #2 (1600 rounds, 200 nodes), iAFTA delayed the FND by 12.1 % (1012.4 vs. 903.9 rounds) and 25.3 % (1096.2 vs. 875.2 rounds), respectively, while maintaining partial functionality beyond round 1600. BWOA-V fully degraded by round 1250.7. In Scenario #3 (100x100m area), all nodes survived under iAFTA, whereas BWOA-V recorded an FND at 1137.1 rounds. Furthermore, iAFTA reduced AEC by 93.3 % at round 500 (0.0138 J vs. 0.2056 J for BWOA-V) and ensured robust fault tolerance with only 11 CH replacements, balancing residual energy among CHs in later rounds. While these results validate iAFTA’s scalability and resilience for large-scale WSN applications, certain limitations were identified during testing under high-stress conditions in Scenarios #4 and #5 (15,000 rounds). In Scenario #4, the LND metric was undefined across all runs due to simulation termination before complete network depletion, limiting the ability to fully assess overall network lifespan but confirming that iAFTA prevents simultaneous energy exhaustion across nodes. In Scenario #5, a high node density of 900 nodes resulted in an average of 4074 CH replacements, reflecting robust energy redistribution but also exposing challenges in achieving balanced energy depletion across all nodes.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

iAgent fault-tolerance approach (iAFTA) based on optimization algorithms in wireless sensor networks

  • 期刊论文

  • 发表日期：

    2025年6月18日

查看全文
  • Mouna Ktari,
  • Raïda Ktari,
  • Yassine Khemakhem,
  • Ahmed Hadj Kacem

摘要

Wireless sensor networks (WSNs) face persistent challenges in energy efficiency and fault tolerance due to limited node battery life. This paper introduces a novel multi-goal intelligent agent approach (iAFTA) for static WSNs, where intelligent agents (iAgents) proactively monitor cluster head (CH) energy, detect failures, and dynamically elect replacements based on energy and proximity. The process integrates fault tolerance to prevent cascading failures while minimizing energy consumption through adaptive minimum spanning tree (MST) reconfiguration. The proposed iAFTA approach is evaluated across five distinct scenarios using key metrics such as first node dead (FND), half node dead (HND), last node dead (LND), average energy consumption (AEC), and CH replacement rate. In addition, fault-tolerance metrics, including fault detection accuracy, detection time, and energy overhead, are evaluated. Results show that iAFTA achieves 100 % fault detection accuracy, immediate detection and recovery (average detection time of 1.0 round), and low-energy overhead during fault handling, even in large-scale scenarios. Results demonstrate significant improvements over existing protocols like BWOA-V. In Scenarios #1 and #2 (1600 rounds, 200 nodes), iAFTA delayed the FND by 12.1 % (1012.4 vs. 903.9 rounds) and 25.3 % (1096.2 vs. 875.2 rounds), respectively, while maintaining partial functionality beyond round 1600. BWOA-V fully degraded by round 1250.7. In Scenario #3 (100x100m area), all nodes survived under iAFTA, whereas BWOA-V recorded an FND at 1137.1 rounds. Furthermore, iAFTA reduced AEC by 93.3 % at round 500 (0.0138 J vs. 0.2056 J for BWOA-V) and ensured robust fault tolerance with only 11 CH replacements, balancing residual energy among CHs in later rounds. While these results validate iAFTA’s scalability and resilience for large-scale WSN applications, certain limitations were identified during testing under high-stress conditions in Scenarios #4 and #5 (15,000 rounds). In Scenario #4, the LND metric was undefined across all runs due to simulation termination before complete network depletion, limiting the ability to fully assess overall network lifespan but confirming that iAFTA prevents simultaneous energy exhaustion across nodes. In Scenario #5, a high node density of 900 nodes resulted in an average of 4074 CH replacements, reflecting robust energy redistribution but also exposing challenges in achieving balanced energy depletion across all nodes.