The isolated toughness variant is a salient parameter for measuring the vulnerability of networks, which is inherently related to fractional factors (used to characterize the feasibility of data transmission). The combination of minimum degree and the corresponding tight bound of isolated toughness variant for fractional factors provide reference standards for network construction. However, previous advances only focused on how to select the optimal parameter criteria from Pareto front, without any suggestion for the construction of specific networks. To overcome this deficiency, this paper proposes an inverse model from $(δ(G),I'(G))$ to $G$, by means of evolutionary computing approach, we propose a novel inverse model to obtain the optimal solutions for candidate graphs. The main procedure is composed of pseudo-greedy acceleration, cross-mutation and diversity enhancement modules. The practicality of the algorithm is verified by means of pilot experiments. The code in this paper is made public on https://github.com/AizhEngHN/Inverse-model-for-network-construction-G-I-G-G.