Fudan University predicts Trump will be the next president. This is from Fudan U’s CCDA (Center for Complex Decision Analysis). They predict Trump will win (60%), based not on polling or structural analysis, but on agent-based computer modelling (ABM). 復旦大學預測川普將成為下一任總統。這是來自復旦大學的（複雜決策分析中心）。他們預測川普將獲勝（60%），不是基於民意調查或結構分析，而是基於基於代理的電腦建模.
These are the first known applications of ABM to election predictions. Past predictions have been shown to be not only correct, but accurate to actual outcomes of elections. 這些是 ABM 在選舉預測中的首次已知應用。過去的預測已被證明不僅是正確的，而且準確地反映了選舉的實際結果.
They take various key factors (individual: age, gender, income, education, occupation, ethnicity, religion; macro factors: economic growth, gini coefficient, unemployment; candidate attributes: incumbency, age, gender, education, scandals, etc) and do regression analysis (i.e. how much each factor affects an individuals’ candidate choice based past election results). They use the regression analysis to derive a set of rules guiding the voting preference of multiple agent models in the future. These bottom-up agent models are then run again in simulation to see if they can predict the results of past elections. The ones that are the most successful are then run in simulation for forecasting a future election, in this case, 2024. 他們考慮各種關鍵因素（個人：年齡、性別、收入、教育、職業、種族、宗教；宏觀因素：經濟成長、基尼係數、失業率；候選人屬性：在職狀況、年齡、性別、教育、醜聞等）迴歸分析（即每個因素對個人基於過去選舉結果的候選人選擇的影響程度）。他們利用迴歸分析得出一組指導未來多個代理模型的投票偏好的規則。 然後，這些自下而上的代理模型再次在模擬中運行，看看它們是否可以預測過去的選舉結果。 然後，模擬運行最成功的預測來預測未來的選舉，在本例中為 2024 年.
The model is designed to take into account “shock factors”, although we don’t know which. Allegations of North Korea (NK) participation, Trump’s relationship to NK, and the Ukraine war could be an unfactored variable. 該模型的設計考慮了“衝擊因素”，儘管我們不知道是哪些因素。對北韓參與的指控、川普與北韓的關係以及烏克蘭戰爭可能是一個未考慮的變數.
It also assumes no outside (e.g. Deep State or other) intervention (i.e. assassination, fraud, dirty tricks, lawfare), 它還假設沒有外部（例如深州或其他）幹預（即暗殺、欺詐、骯髒伎倆、法律）
Article on Fudan U’s ABM Methodology for predicting elections 復旦大學預測選舉的 ABM 方法論文章
https://pmc.ncbi.nlm.nih.gov/articles/PMC9246136/
Good intro text on ABMs & complexity (contains some math & python programming but concepts can be understood without) 關於 ABM 和複雜性的良好介紹文本（包含一些數學和 Python 編程，但概念可以不理解）
Business Blog 