Abstract: China as a big steel producer, sufficient iron ore is a necessary condition for steel mills to ensure continuous production. On the one hand, due to low grade, difficult mining and the complexity of mineral processing of iron ore resources at home, it is necessary to import a large number of foreign high-quality iron ore to make up for domestic demand. However, foreign high-quality iron ore is monopolized by a few countries (such as Australia, Brazil, South Africa, India), which leads to high cost of procurement logistics of iron ore. On the other hand, a large number of iron ore require the steel mill to invest large amounts of money for expanding the yard. As a logistics node for overseas trade, the port cannot be ignored in supply chain of the imported iron ore. But existing shoreline resources of the port are not fully utilized. "Port before Factory", as its name implies, "the front is the port, the rear is the steel mill", means that the port and the steel mill sign a long-term contract. Under this mode, after the raw materials purchased by the steel mill are shipped to the unloading port, they are stored in the port yard. According to the production instructions of the steel mill, the port sends a certain amount of raw materials to the steel-making furnace of the steel mill in a certain transportation method on time. Or the products to be sold produced by the steel mill are directly stored in the port yard, where products are shipped according to the received sales instructions of the steel mill, thereby forming a "closed-loop" operation. The "Port before Factory" mode links the port with the steel mill's procurement, production and sales, and promotes the deep horizontal cooperation between the port and steel mill.
Therefore, based on the "Port before Factory" mode, this paper constructs a combination optimization model with the selection of ship type, supplier and unloading port as variables to minimize the cost of iron ore procurement logistics. This model has the characteristics of multi-variable, multi-constraint and non-linearity. It is a large-scale combinatorial optimization model. Because the immune genetic algorithm(IGA) has wide applicability in solving such problems, it not only retains the search characteristics of genetic algorithms, but also utilizes the adaptive characteristics of the multi-mechanism of immune algorithm to solve the optimal solution of multi-objective functions. To a large extent, it avoids "premature convergence" to local extremum. Therefore, the IGA is used to solve the model in this paper.
In order to verify the superiority of the "Port before Factory" mode, relevant data such as ore suppliers, docking ports, ship types, steel mills, etc. were collected. Substituting these data into the model, the cost value obtained by usingIGA and MATLABis 44.1 billion yuan. In the traditional mode, the whole process cost of procurement under the same conditions is 52.2 billion. The former saved 15.5% of the cost compared to the latter. In addition, the combination optimization scheme shows that: a) Australia and Brazil are China's major iron ore suppliers, accounting for about 90% of orders, which is determined by their supply capacity, distance advantage and large ships; b) Port B and port C respectively undertake 57.4% and 42.6% of iron ore, of which Australia prefers to choose port C, Brazil mainly berths in port B; (c) According to port natural and facilities conditions, Panamax and Capesize small vessels are more likely to berth in port C, while Vloc and Valemax large vessels are more likely to choose port B. To sum up, it provides a theoretical basis for the implementation of the "Port before Factory" mode.
Today, in the era of "large wharf, large chemical industry and large steel", the successful practice of "Port before Factory" mode makes "back factory" no longer limited to steel enterprises, and can be extended to "large chemical industry, large oil refining" and other large users and large systems, so as to promote the win-win situation of industry and port, share social responsibility-environmental protection, and implement the goal of sustainable development.

Key words: "Port before Factory", closed-loop, the cost of procurement logistics, ship type, unloading port