In this paper, a global soybean trade network (GSTN) is constructed based on the international trade data from 1987 to 2020. The results are as follow: (1) From 1987 to 2020, the scale of the network continued to expand, the number of trading countries continued to increase, and the trade links between countries gradually deepened. (2) The power function curve of the node degree distribution shows that the GSTN satisfies the power law distribution, which fits to be a scale-free network. (3) The global soybean export market is highly concentrated. The United States, Brazil and Argentina account for more than 80% of global soybean exports. China is the world’s largest soybean importer. (4) The control ability of core countries in the GSTN has gradually declined, which is shown from that the importance of nodes has gradually weakened, while the trade between peripheral countries is gradually increasing, showing the phenomenon of trade multilateralism. (5) South America has become the global soybean export centre, and Asia has become the global soybean import centre. The trade status of North America has declined, and the overall strength of Africa and Oceania is the weakest.

Grain is the material basis for human survival and development, wheat is one of the three significant grains, and the robustness of the wheat trade has seriously affected world grain security. In the paper, we analyze the robustness of the International Wheat Trade Network (IWTN) from three perspectives: global, regional, and national. We quantify the topological properties and evolution characteristics of IWTN from 1987 to 2020, and analyze the stability of IWTN using random attacks and target attacks. The import satisfaction of major wheat trading countries and trade communities is quantified through import stability indicators. Firstly, the results show that IWTN satisfies the scale-free property and that global wheat trade transactions continue to increase. Secondly, IWTN under random attack forms a “convex” curve regarding network efficiency. It shows that IWTN has better resistance to external shocks, and higher robustness. However, under the targeted attack, the network collapses when the node failure rate reaches 40%. Regardless of the attack mode, the decreasing trend of network efficiency and connectivity is the same in different years as the number of failed nodes gradually increases, and no year shows outstanding stability. Thirdly, different communities in the wheat trade are more satisfied with the wheat imports of their internal countries, and most of the wheat trade occurs within the same trading community. Fourthly, among the top five global wheat importers in 2020, Italy and Turkey have diversified sources of imports, but trade imports are relatively concentrated. In contrast, China, Egypt and Indonesia have relatively balanced wheat trade imports from their respective trading countries and are less at risk of being affected in the event of supply risks.

The function of the network is mainly related to its composition and structure. With the arrival of the big data era, the network congestion has become more and more serious. Many networks are proved to either have multilayer nature in themselves or interdepend with other networks. We focus on the multilayer network model formed by different speeds layers and study traffic dynamics on four different structure multilayer network models based on random and scale-free networks. It is found that the low-speed layer has a greater impact on the network traffic, and this layer with homogeneous network structure can better avoid congestion. Moreover, simulation results show that high-speed and low-speed layers with heterogeneity structure is easy to make the multilayer networks occur traffic congestion while the homogeneity structure is beneficial to obtain large traffic capacity.