Recently, the number of deteriorating bridges has drastically increased. Furthermore, tight maintenance budgets are cut down, imposing escalating adverse implications on the safety of bridges. This state of affairs entails the development of decision support systems for the effective management of bridges within the allocated budget. As such, this study introduces an invasive weed optimization-based fuzzy decision-making framework designated for bridge intervention prioritization in both element and network levels. The proposed decision-making platform encompasses three main tiers. The first tier is an optimized fuzzy analytical network process model that aims at computing the weighting vector of the bridge defects, namely corrosion, delamination, cracking, spalling and scaling. In this model, a genetic algorithm optimization model is formulated to improve the consistencies of judgment matrices through circumventing the imprecisions encountered by the classical judgment assignment. The second tier encompasses establishing an integrated bridge deck condition assessment model capitalizing on ground-penetrating radar and inspection reports. In it, the severities of the bridge defects are demonstrated in the form of fuzzy membership functions to address the inherent uncertainties of inspection. Subsequently, a variable-length invasive weed optimization model is structured to automatically calibrate the fuzzy membership functions. The third model is designed for structuring a bridge maintenance decision-making strategy stepping on the integrated condition index. The capabilities of the proposed framework were validated through several levels of comparisons. For instance, it significantly outperformed some of the current condition assessment models. Additionally, it inferred that the thresholds separating the four categories of the integrated bridge deck condition index are 75.651, 67.769 and 60.318.

中文翻译：

---

一种基于侵入性杂草优化的模糊决策框架，用于元素和网络级别的桥梁干预优先级

最近，损坏的桥梁数量急剧增加。此外，紧张的维护预算被削减，对桥梁安全的不利影响不断升级。这种情况需要开发决策支持系统，以便在分配的预算内有效管理桥梁。因此，本研究引入了一种基于侵入性杂草优化的模糊决策框架，用于在元素和网络级别进行桥梁干预优先级排序。拟议的决策平台包括三个主要层次。第一层是优化的模糊分析网络过程模型，旨在计算桥梁缺陷的权重向量，即腐蚀、分层、开裂、剥落和结垢。在这个模型中，建立遗传算法优化模型，通过规避经典判断分配遇到的不精确性，提高判断矩阵的一致性。第二层包括利用探地雷达和检查报告建立综合桥面状况评估模型。其中，桥梁缺陷的严重性以模糊隶属函数的形式展示，以解决检查的固有不确定性。随后，构建了可变长度入侵杂草优化模型以自动校准模糊隶属函数。第三个模型是为构建基于综合状况指标的桥梁维护决策策略而设计的。通过多个级别的比较验证了所提议框架的功能。例如，它显着优于当前的一些状况评估模型。此外，它推断综合桥面状况指数四类的阈值分别为75.651、67.769和60.318。