一句话：提出“设计多元宇宙”模型，把设计演化刻画为上下文敏感的非确定性迁移系统，并正交分离用于一致性推理的观测链接理论。
模型驱动系统工程（MBSE）目前缺少一种把设计演化——及其固有的分支、返工与上下文敏感性——当作一等公民的形式模型。现有做法把丰富的设计轨迹降格为仓库中的制品，从而丢失了设计决策本身的语义。本文呼吁建立设计演化的科学模型，并提出“设计多元宇宙”（Design Multiverse）：它将设计动态刻画为一个上下文敏感的非确定性迁移系统，同时正交地分离出观测链接理论（如 conformance、refinement），以支持一致性推理。作者通过协同演化（co-evolution）场景展示了该模型的解释力，并进一步提出以设计多元宇宙实验（DMX）为支撑的研究议程，使该模型可被实证检验，为可扩展且具历史感知能力的建模环境铺路。
Model-Based Systems Engineering lacks a formal model that treats design evolution, with its inherent branching, rework, and context-sensitivity, as a first-class citizen. Current approaches reduce rich design trajectories to repository artifacts, losing the semantics of design decisions. This paper calls for a scientific model for design evolution and introduces the Design Multiverse that frames design dynamics as a context-sensitive, nondeterministic transition system, while orthogonally separating observational link theories (e.g., conformance, refinement) for consistency reasoning. We demonstrate the model’s explanatory power through co-evolution. Furthermore, we outline a research agenda grounded by Design Multiverse Experiments (DMX), which enable empirical validation and pave the way for scalable, history-aware modeling environments.
@inproceedings{paperbot12030,
title = {The Design Multiverse: A Scientific Model for Design Evolution and Co-evolution},
author = {Ciprian Teodorov and Joeri Exelmans and Salvador Martínez and Sylvain Guérin and Hans Vangheluwe},
booktitle = {ICSE},
year = {2026},
doi = {10.1145/3786582.3786830}
}