After 30 years of active development, mechanistic models of the reproductive process nearly stopped attracting scholarly interest in the early 1980s. In the following decades, fertility research continued to thrive, relying on solid descriptive work and detailed analysis of micro-level data. The absence of systematic modelling efforts, however, has also made the field more fragmented, with empirical research, theory building, and forecasting advancing along largely disconnected channels. In this paper we outline some of the drivers of this process, from the popularization of user-friendly statistical software to the limitations of early family building models. We then describe a series of developments in computational modelling and statistical computing that can contribute to the emergence of a new generation of mechanistic models. Finally, we introduce a concrete example of this new kind of model, and show how they can be used to formulate and test theories coherently and make informed projections.

经过 30 年的积极发展，生殖过程的机械模型在 80 年代初期几乎不再引起学术界的兴趣。在接下来的几十年里，生育研究继续蓬勃发展，依赖于扎实的描述工作和对微观数据的详细分析。然而，缺乏系统的建模工作也使得该领域更加分散，实证研究、理论构建和预测沿着基本上不连贯的渠道推进。在本文中，我们概述了这一过程的一些驱动因素，从用户友好的统计软件的普及到早期家庭建设模型的局限性。然后，我们描述了计算建模和统计计算方面的一系列发展，这些发展可能有助于新一代机械模型的出现。最后，