Modeling the warring states period: History dynamics of initial unified empire in China (475 BC to 221 BC)

•ABM is applied to model the distinctive unification pattern of ancient China.•The optimal solution can precisely back-calculate actual history.•Alliances and wars influence the Unity pattern, and wars have a greater impact.•Marginal diminishing and asymmetric effects reveal the impact of war cost. For the evolution of civilization, we focus on how the Unity pattern was formed in ancient China. Apart from existing case studies and mathematical models, we use agent-based modeling (ABM) to explore the initial unity process. Before this, we have a long chaotic dynasty, which is the Eastern Zhou. It has two periods, such as the Spring and Autumn period (770 BC-475 BC) and Warring States period (475 BC-221 BC). The unity outcome was achieved by interactions or interplays (wars, alliances, and mergers) within autonomous states (agents). We have built a general ABM model for this whole dynasty. However, the mechanisms during two periods are different in the history, and simulation outcomes are also different. Hence, we need to investigate them one by one. Based on ABM of the first period, we further explore the history dynamics of the Warring States period (475 BC-221 BC). For mechanism designs, we revised the war costs, alliance rules, and war behaviors, according to the detailed history. For verification standards, we have added real number of wars and alliances in the history, to double-check our general model. It seems that real history data can be well matched by our simulations, and the best-fitted solution can be found. Based on simulations and counterfactual experiments, we can obtain accurate effects of related factors. It seems that: the bellicosity will reduce the duration and accelerates the unification process; the alliance propensity gradually increases or prolong the duration with more wars and alliances. Moreover, the bellicosity (war) has the peak effect (around 0.2) on the frequencies of wars and alliances. The alliance propensity behavior has linear effect, and it always increases the wars. Generally, the bellicosity (war) plays the bigger role than the alliance. If more agents are changing alliance partners (less reliable of the alliance network), the unification process can be fastened. The opportunistic propensity has the peak effect (around 0.7) on the frequencies of wars and alliances. Marginal diminishing and asymmetric effects of war costs can be revealed as well.