Prediction of global value chain based on cognitive neural network -Take Chinese automobile industry as an example

3.1 The Introduction to Fuzzy Theory

When some enterprises in the value chain carry out production according to the parameters (standards or rules) set by other subjects, the problem of governance arises. These parameters include what to produce (definition of products), how to produce (i.e. definition of producing process, including elements such as technique, quality, labor and environmental standards), when to produce, how much to produce, and price. On the left of point A is the production field, and on the right is the circulation field. Both fields have a circulation process, and the minimum value-added link is the link of the vehicle assembly, as shown in Figure 3.

Figure 3

Value distribution curve of automobile industry chain

3.2 Cognitive Artificial Neural Network

Kaplinsky & Morris (2001) talked about cluster upgrading from the perspective of value chain upgrading and thought there are mainly four types of cluster upgrading: process upgrading, production upgrading, functional upgrading and chain upgrading. Process upgrading refers to increasing input-output level by reorganizing product systems or introducing high technologies; production upgrading refers to the development of new products, adopting more complex product lines and improving quality faster than competitors; functional upgrading refers to accepting new functions or giving up old functions to improve skills; chain upgrading refers to shifting to new and relevant industrial value chains of high value (see Table 1) 67.

3.3 The integration of Fuzzy System and Cognitive Artificial Neural Network

In many links of a value chain, not every link creates equal value. The global economy is not bound by the standards and rules set by the free market alliances and intergovernmental organizations, but by the rules and standards set by the leading companies in the global value chain of the industry, and on this basis, the global value is proposed. The concept of chain governance: Governance of global value chains refers to the realization of non-market coordination between different economic activities and different links in the value chain through the relationship arrangement and institutional mechanism between companies in the global value chain.

One of the basic ideas in the study of global value chain is that among numerous value links

in the whole global value chain of a certain industry, it’s not every link that creates value, and according to the value-added capacity of each link, the whole value chain can be divided into several links or segments. In other words, the whole chain of values is characterized by a hierarchy of values [8]. The high value-added value link is generally the core link of the global value chain, and the global governance rules of the whole value chain are also determined by these core links. According to the level of value-added ability, there exists a strict hierarchy system of comparative advantages in different regions. The process of matching the value hierarchy system of global value chain with the global comparison hierarchy in different regions is the process of vertical separation and spatial reconstruction of value links in global value chain. The hierarchy system of local industrial clusters on the same value chain in the real world is determined by the hierarchy system of value links. According to geographic spatial relationship, it can be divided into global value chain, transnational value chain, domestic value chain and local value chain, as shown in Figure 4.

Figure 4

The upgrading approach of China’s automobile industry in the global value chain

4.1 Cognitive Artificial Neural Networks algorithm

Humphrey and Schmitz’s method on value chain classification has increased the way that industrial clusters are embedded in the global value chain. Corresponding to the four types of their chain governance models, there are four embedding modes of local industrial clusters: bureaucracy embedding, which refers to one behavior subject is completely controlled by the other, or completely control the other, for example, being embedded into the global value chain by acquisition or being acquired; market

embedding, which refers to that the economic actors are embedded in the global value chain through the direct purchase and sale of various goods or services, and the parties are fully symmetric in power and only need to trade through a price mechanism, such as the direct sale of goods to overseas markets by means of international trade; network embedding, which refers to one actor is embedded in the global value chain with the complementary advantages needed by other actors, and all parties are symmetric in power, but often coordinate some activities by non-price mechanism; quasi-bureaucracy embedding, which refers to one actor becomes the dominant person in the global value chain by virtue of some advantages, while the other actors are in a subordinate position, such as embedding the global value chain through OEM, ODM, etc. and quasi-bureaucracy embedding includes two types of embedding such as buyer driving embedding and producer driving embedding.

Artificial neural network is a statistical tool that simulates human brain for thinking and learning. BP neural network is the most widely used artificial neural network. BP algorithm is a three-layer topology including input layer, output layer and hidden layer. There is a mentoring learning algorithm. The basic principle is error back propagation and information forward propagation. The operation process is: training the network with the sample, continuously changing the connection weight between the neurons, and finally making the difference between the expected output value and the actual output value within a preset error range, thereby stopping the training, and the process of storing the network connection weights and completing the training.

Back propagation is a multilayer feed forward network trained by error back propagation algorithm and is one of the most widely used neural network models. BP network can learn and store a large number of input-output pattern mapping without revealing the mathematical equation describing the mapping in advance. Its learning rule is to use the steepest descent method to adjust the weights and thresholds of the network continuously through the backward propagation, so as to minimize the sum of the square errors of the network. The model topology of BP neural network includes input layer, hidden layer and output layer. The cognitive artificial neural networks model is shown in Figure 5.

Figure 5

Cognitive Artificial Neural Networks algorithm

4.2 Simulation Calculation

Although the structure of the industry can be changed, it’s not good to be excessively dependent on changes in the structure of the industry to choose their new value chain and new customers before entering the industry. The change in the structure of the industry not only needs strong strength, but also works like a double-edged sword, since it can improve the profitability of the structure of the industry, but it can easily destroy this profitability. For example, a new product design that removes barriers to invasion or exacerbates competitiveness could destroy the industry’s foundation for long-term profitability, even though the promoters of the new product would temporarily make huge profits. In general, enterprises often neglect the long-term impact on the industry structure when they make strategic decisions, and they can only see the temporary improvement of their competitive position due to temporary success, but cannot foresee the reaction of competitors, which may have bad consequences. If main competitors follow suit, the profitability of the industry structure will be greatly weakened, so that all enterprises in the industry are suffering from it; however, even if all enterprises in the industry pursue a modest and passive market strategy, the competition in the industry will not necessarily be completely alleviated, because this will induce new competitors to enter, and then stir up competition. So, the more intense the competition in the industry is, the more helpful it can prevent new competitors from entering.

4.3 Data selection

Competitive advantage is the core of enterprise performance in competitive market. There are two basic forms of competitive advantage: cost leadership and differentiation. The competitive advantage comes from many separate activities carried out by enterprises in the process of design, production, marketing, delivery and other auxiliary processes. Each of these activities contributes to the relative cost position of the enterprise and lays the foundation for differentiation. The difference between competitors’ value chain is the key source of the competitive advantage. As a strategic tool, value chain plays an important role in analyzing relative cost position, differentiation and obtaining competitive advantage. Cultivating core competitiveness achieves sustainable development of enterprises. Through analyzing the relationship between the internal and external value chain of Laigang Group and the competitive advantage, we can see that Laigang Group has certain scale economic advantage in the industry, but it has no obvious advantage in variety, geographical location, resources and enterprise management. Therefore, Laigang Group should optimize the allocation of resources from a strategic perspective, foster the core competitiveness of enterprises, and achieve sustainable development. It should strengthen strategic management and carry out strategic positioning. In terms of overall competitive strategy, Laigang Group should mainly adhere to the leading cost, and strive to improve the differentiated advantages of such high value-added products as special steel, plates and strips. In order to integrate the cost advantage and the difference advantage, the whole value activity complex must be reconfigured. Table 2 shows the Auto Industry International Share Index from 2004 to 2013, Comparison of Auto industry international share index from 2004 to 2013 as shown in Figure 6.

Figure 6

Comparison of Auto industry international share index from 2004 to 2013