Underwater tank cleaning using robotic method is very crucial due to the concern on the diver’s safety in undisrupted water supply operation. A Remotely Operated Underwater Vehicles (ROV) used in the tank cleaning operation however, suffers from a high operational cost due to the lack of systematic operator guidance in robot maneuvering. This paper presents a multi-objective approach in designing a Decision Support System (DSS) for underwater cleaning robot. To explore all feasible path, the path alternatives for every cleaning point in the tank is found using Probabilistic Roadmap (PRM). Then, an optimized sequential route are identified using Non-Dominated Sorting Genetic Algorithm using Reference Point Based (NSGA-III). Several objectives such as path length and routing angle are considered to be optimized, while ensuring constraints such as similar deployment point, maximum daily time limit and cable entanglement. To measure the quality of the proposed solution, comparisons have been done based on performance of NSGA-III with Non-Dominated Sorting Genetic Algorithm II (NSGA-II) and Multi-Objective Particle Swarm Optimization (MOPSO) by considering the C-Metric value, execution time and estimated cleaning duration. In addition, comparisons with conventional path by human operator is also conducted to validate the significance of DSS application in underwater tank cleaning. Results have shown that NSGA-III has superiorities with an improvement of 11.11% in cleaning time as compared to NSGA-II and 5.12% improvement compared to MOPSO.

由于担心潜水员在不间断的供水操作中的安全，使用机器人方法清洗水下水箱非常重要。但是，由于缺乏机器人操纵系统的操作指导，用于罐清洗操作的遥控水下航行器（ROV）遭受了高昂的运行成本。本文提出了一种设计水下清洁机器人决策支持系统（DSS）的多目标方法。为了探索所有可行的路径，可使用概率路线图（PRM）找到水箱中每个清洁点的路径替代方案。然后，使用基于参考点的非支配排序遗传算法（NSGA-III）确定优化的顺序路线。路径长度和走线角度等几个目标被认为是最优化的，同时确保诸如类似部署点，最大每日时间限制和电缆缠绕之类的限制。为了衡量所提出解决方案的质量，已经基于NSGA-III与非主导排序遗传算法II（NSGA-II）和多目标粒子群优化（MOPSO）的性能进行了比较，并考虑了C-Metric值，执行时间和估计的清洁时间。此外，还与操作人员进行了常规路径比较，以验证DSS在水下水箱清洁中的应用意义。结果表明，NSGA-III具有优势，与NSGA-II相比清洁时间提高了11.11％，与MOPSO相比提高了5.12％。基于NSGA-III的性能与非主导排序遗传算法II（NSGA-II）和多目标粒子群优化（MOPSO）进行了比较，考虑了C度量值，执行时间和估计的清洁时间。此外，还进行了人工操作人员与常规路径的比较，以验证DSS在水下水箱清洁中的应用意义。结果表明，NSGA-III具有优势，与NSGA-II相比清洁时间提高了11.11％，与MOPSO相比提高了5.12％。基于NSGA-III的性能与非主导排序遗传算法II（NSGA-II）和多目标粒子群优化（MOPSO）进行了比较，考虑了C度量值，执行时间和估计的清洁时间。此外，还与操作人员进行了常规路径比较，以验证DSS在水下水箱清洁中的应用意义。结果表明，NSGA-III具有优势，与NSGA-II相比清洁时间提高了11.11％，与MOPSO相比提高了5.12％。还与操作员与常规路径进行了比较，以验证DSS在水下水箱清洁中的应用意义。结果表明，NSGA-III具有优势，与NSGA-II相比清洁时间提高了11.11％，与MOPSO相比提高了5.12％。还与操作员与常规路径进行了比较，以验证DSS在水下水箱清洁中的应用意义。结果表明，NSGA-III具有优势，与NSGA-II相比清洁时间提高了11.11％，与MOPSO相比提高了5.12％。