复杂工程决策中关联个体动态有限妥协的最小调整共识方法

A Minimum-Adjustment Consensus Method with Dynamic Limited Compromise of Associated Individuals for Complex Engineering Decision-Making

摘要: 复杂工程研发与方案评审通常需要多领域个体协同参与，个体间潜在关联关系及妥协行为的动态变化会直接影响工程方案的共识形成与决策稳定性。针对关联关系未知以及妥协状态难以动态刻画的问题，本文提出一种考虑关联个体动态有限妥协的最小调整共识方法。首先，根据意见距离和有界置信识别潜在关联关系，并结合意见相似程度确定关联权重，由相关联成员意见形成个体关联意见；随后，根据个体自身意见与关联意见之间的差异学习初始妥协程度。当初始关联关系和妥协状态不能满足共识要求时，通过调整关联关系改变个体关联意见，并据此进一步调整妥协程度，同时以最小调整控制相对于初始状态的变化。最后，以抗癌候选药物筛选为案例进行验证。结果表明，所提出方法能够刻画关联关系变化对个体妥协行为的影响，并在控制调整幅度的同时促进群体共识达成。

Abstract: Complex engineering R&D and scheme evaluation often require coordinated judgments from individuals with heterogeneous expertise. In such settings, latent association relationships and dynamically changing compromise behavior can directly affect consensus formation and decision stability. To address unknown association relationships and dynamic compromise states, this study proposes a minimum-adjustment consensus method with dynamic limited compromise. Potential associations are first identified using opinion distance and bounded confidence, and association weights are determined by opinion similarity to form individual reference opinions. Initial compromise degrees are then learned from the differences between individual and reference opinions. If the initial state fails to meet the consensus threshold, association weights and compromise degrees are further adjusted while minimizing deviations from the initial state. An anticancer drug screening case shows that the proposed method can capture the effect of changing associations on compromise behavior and improve consensus with limited adjustment.