基于NK模型的重大工程决策关联复杂性分析

Interrelated Complexity Analysis of Major Engineering Project Decision-making Based on NK Model

摘要:

关联复杂性是重大工程决策过程中的重要特征，但相关研究较多在传统的项目管理框架内展开，显现出一定片面性和局限性。为此，研究将NK模型引入重大工程决策关联复杂性分析，通过梳理国内外相关研究，对重大工程决策关联复杂性作出定义并明确其具有的关联性、非线性、自适应性等特征内涵；从决策环境、主体和目标出发，分析重大工程决策关联复杂性并构建理论模型；以港珠澳大桥为例，提出重大工程决策关联复杂性带来的4个主要问题：决策结构、决策权力、主导决策和内外部关联性问题，并引入模块化影响矩阵、适应性游走方式、基因型搜索方式和参数K、C仿真等NK模型所特有的问题解决方法。研究认为，基于NK模型的问题解决方法能够较好地契合重大工程决策过程中各决策点自下而上彼此关联、交互影响的特征，为后续相关研究提供了模型框架和理论参考。

Abstract:

Interrelated complexity is a critical feature in the decision-making process of major engineering projects. However, most related research has been conducted within the traditional project management framework, revealing certain partialities and limitations. To address this, the study introduces the NK model into the analysis of interrelated complexity in major engineering decision-making. By reviewing relevant research both domestically and internationally, the study defines interrelated complexity in major engineering decisions and clarifies its inherent characteristics, such as interrelatedness, nonlinearity, and adaptability. Starting from the decision-making environment, subjects, and objectives, the study analyzes the interrelated complexity of major engineering decisions and constructs a theoretical model. Using the Hong Kong-Zhuhai-Macao Bridge as an example, the study identifies four major issues brought about by interrelated complexity in major engineering decisions: decision structure, decision authority, leading decision-making, and internal and external interrelatedness. The study introduces solutions specific to the NK model, such as modular impact matrices, adaptive walks, genotypic search methods, and simulation parameters K and C. The study concludes that the NK model-based problem-solving methods can effectively align with the bottom-up, interrelated, and interactive features of decision points in the decision-making process of major engineering projects, providing a model framework and theoretical reference for subsequent related research.