Compressive arch action (CAA) is one of the favorable resistance mechanisms at low deflections against progressive collapse of reinforced concrete (RC) frame buildings. In this study, Park & Gamble model for calculating the CAA capacity of RC frame structures is modified by incorporating a compressive stress block for high strength concrete (HSC). Meanwhile, a comprehensive database of RC specimens is collated and the modified model is applied on the database to determine the peak deflection (δ) values at CAA by using an iterative approach. The δ values obtained are then modelled by using distinctive features of gene expression programming (GEP) to develop a relationship for δ as a function of height (h) and net span length of the two beams after the column removal (L). The proposed model is validated against a wide range of published results and compared with other models available in literature. The comparison shows that the proposed modified Park & Gamble model can accurately predict the CAA capacity of the specimens. Moreover, the developed model is easy-to-use and skips the oversimplified assumption proposed in the Park & Gamble model to calculate δ (i.e. δ = 0.5*depth of the specimen).

压缩拱作用（CAA）是在低挠度下抵抗钢筋混凝土（RC）框架建筑物逐渐倒塌的有利抵抗机制之一。在这项研究中，通过合并用于高强度混凝土（HSC）的压应力块，修改了用于计算RC框架结构的CAA能力的Park＆Gamble模型。同时，整理了一个完整的钢筋混凝土试样数据库，并将改进的模型应用于该数据库，以使用迭代方法确定CAA处的峰值挠度（δ）值。然后，使用基因表达编程（GEP）的独特特征对获得的δ值进行建模，以建立δ与高度（h）和拆除柱子后两根梁的净跨度（L）。所提出的模型已针对广泛的公开结果进行了验证，并与文献中的其他模型进行了比较。比较表明，提出的改进的Park＆Gamble模型可以准确地预测样品的CAA能力。此外，所开发的模型易于使用，并且跳过了Park＆Gamble模型中提出的过分简化的假设来计算δ（即δ  = 0.5 *样本深度）。