本文解的埃奇沃思模型描述了市场的不稳定。哈罗德霍特林在1929年质疑这一观点；他认为价格或产量不稳定的寡头垄断的本质特征。经济学在研究市场结构对资源配置效率的影响时，根据市场每一个企业可以影响其产品的价格，将市场分为完全竞争市场和不完全竞争市场，垄断市场。网络经济给传统经济中的垄断与竞争概念赋予了新的含义。在网络经济中，竞争会导致垄断的出现。垄断的出现并不能消除竞争的状态。在网络经济条件下，高固定成本、低边际成本、正反馈机制、高转移成本和“锁定”效应成为产品垄断的因素。同时，网络经济中的垄断态势不稳定。创新也是网络经济的特征。因此，竞争无处不在。

The explanations of the Edgeworth model describe the instability in the market for only two sellers. Harold Hotelling challenged this view in 1929; he argued that price or output instability was not an essential feature of an oligopoly. Economics in the study of market structure on the efficiency of resource allocation, according to the market each enterprise can affect the price of their products, the market is divided into a fully competitive market and imperfect competitive market, monopoly market. The network economy has given new meaning to the concept of monopoly and competition in the traditional economy. In the network economy, competition will lead to the emergence of monopoly. The emergence of monopoly cannot eliminate the state of competition. In the network economy, the characteristics of high fixed cost, low marginal cost, positive feedback mechanism, high transfer cost and "lock" effect of the products become the factors of monopoly. At the same time, the monopoly situation in the network economy is unstable. Innovation is also the characteristics of the network economy. As a result, competition is ubiquitous and intensified. The network economy tends to monopolize the market mainly in the following reasons: the inherent characteristics of network products. Its characteristics include: the economies of scale of network products, the externalities of network products, the locking effect of network products and the standardization of network products, copyright protection of network products and mergers and acquisitions in the network economy. The spatial model has received sustained attention from economists over the past few decades.An example of UHM is provided by a television broadcast time or a bus / air schedule. In a timetable game, as in the example above, consumers may be more likely to wait for a period of time relative to the ideal time, rather than earlier arriving to get the service. From the above analysis, we can see that product differentiation will reduce the price elasticity of market demand. Under the same quality conditions, the greater the difference between similar products, the smaller the price elasticity of market demand, the greater the room for enterprises to adjust the price. The product price strategy taken by enterprises is actually a form of market barriers, so as to obtain competitive strategy. To meet the needs of differentiated consumers at the same time, play a set to prevent competitors or potential competitors obstacles. On the other hand, product differentiation will weaken the intensity of price competition. The higher the degree of differentiation of similar products, such as product sales area positioning value of the difference between the greater, the buyer of the product the higher the degree of subjective preferences, at this point the buyer to bear the greater the cost of mobility, while the impact of market prices on consumer demand, the smaller. Therefore, enterprises reduce the price to attract consumers to the less meaningful. In the above Hotelling line segment model, we are to product sales area and brand to discuss the problem of differentiation. In addition to the above two differentiated features, the product also has product quality, product appearance, packaging, market image, compatibility and many other differentiated features. The same conclusion can be obtained if these features are incorporated into the Hotelling line segment model for analysis.

Suppose a linear market with length 1. Consumers in the market evenly distributed. Let x∈ [0,1] denote the location of each consumer. There are two companies, Company A and Company B, whose locations are identified by a and b, respectively. Let us use A to denote the firms in the equilibrium on the left and B the firms on the right. I leave the traditional two-way Hotelling model, assuming that each company can only sell to its left in the consumer. Each firm is produced at a fixed marginal cost, normalized to zero. The fixed cost is zero, but the firm pays the transportation costs and moves the goods from the factory to the consumer(Eaton &Tweedle, 2012). It is reasonable to hypothesis costs and others of linear transportation. That is to say, in order to deliver one unit of product from plant a (accordingly b) to consumer x, company A (or B) pays the transportation cost equal to: t | ax | (corresponding t | bx |. Where t is the (strictly positive) unit transportation cost. The company sets the price for a particular location in the Bertrand game, and sets the number of specific places in the Cournot game. Thus, the inverse demand function is: px = 1 - Qx, where Qx is the total amount provided by the firm at position x Qx ≡ q A(.) + q B(.)). I assume that t ≤ 12: This condition guarantees that there is no local monopoly in the area where both firms can serve, and that none of the locations have a positive amount in the equilibrium. This hypothesis is the standard in the literature on space price discrimination.

In a two-stage game where there is a balance of positions where the company simultaneously selects where to locate, the second stage also selects the price table 4. In this section and below, the sub-game Nash equilibrium concept is used to solve the game (Biscaia&Mota, 2013).

Note that firm A’s equilibrium location depends on the transportation costs. In fact:

In this section, I study the position balance that occurs in a two-stage game in which the company simultaneously selects where to locate and at the same time selects the quantity schedule in the second phase (Tabuchi, 2012). The reason for this difference is as follows. In the Bertrand equilibrium, only Company A serves the consumer on the left, and in Cournot equilibrium, both firms serve consumers on the left side of Company A. Therefore, the strategic effect is not so strong because the rightward movement of firm A is not fully reflected in the reduced markup 12.

Connaught model is by the French economist Anthony Augustine Cournot in 1838 put forward. Is the earliest version of the Nash equilibrium application, the Cournot model is often used as the starting point for oligarchical analysis. The Cournot model is a simple model with only two oligarchs, which is also called a "duopoly model." The model illustrates how the output decisions of firms competing with each other without co-ordination interact with each other to produce a result that lies between competitive equilibrium and monopoly equilibrium. The conclusion of the Cournot model can be easily extended to the case of three or more oligopoly firms.

The Cournot model assumes that there is only two sellers in a product market, and that there is no collusion between them, but how each other knows how each other will act to determine the optimal yield to maximize profit, The model is also called the double headed monopoly theory. The Cournot model analyzes the case of two oligarchs selling zero-cost-of-production for the same product. The Cournot model assumes that only the A and B firms in the market produce and sell the same products and that their production costs are zero; that their common market demand curves are linear and that both firms A and B are accurate To understand the market demand curve; A, B two manufacturers are known to each other in the case of production, each to determine their own to maximize the profits of production, that is, each manufacturer is a negative to their own production to adapt The other has been determined the yield. The above two-headed Cournot model can be generalized. Let the number of oligopolies be m, then the general conclusion can be drawn as follows: The equilibrium output of each oligopoly firm = total market capacity / (m + 1). The industry's total equilibrium output = total market capacity, m / (m + 1). The drawback of the Cournot model is the assumption that the firm does not change the output of its competitors. Bertrand duopoly Model (Bertrand duopoly Model) by the French economist Joseph Bertrand (Joseph Bertrand) was established in 1883. Cournot model and Starkelberg model are the production of the manufacturers as a means of competition is a yield competition model, and Bertrand model is the price competition model. The Bertrand model assumes that when firms formulate their prices, it is assumed that the prices of other firms will not change because of their decisions, and n (for simplicity, n = 2) the products of oligopoly firms are complete substitutes. A, B two companies were the price of P1, P2, marginal cost is equal to C. According to the assumption of the model, A, B, two companies have strong substitutes between the products (completely replaceable, that is, the price is different, the higher prices will be completely sold out), so the consumer choice is the price Low enterprise products; if A, B prices are equal, then the two companies split demand. Therefore, the two companies will compete to cut prices in order to win more customers. When the price down to P1 = P2 = MC, to achieve equilibrium, that Bertrand equilibrium.As long as a competitor exists, the behavior of the firm is the same as in the perfectly competitive market structure, and the price is equal to the marginal cost. According to the Bertrand model, the low price will win the whole market, and the high price will lose the whole market, so the oligarchs will cut each other until the price equals their marginal cost.

In addition, Bournram'sand Cournotby's demand effect is more intense: competition is not intense, the number of competition to increase the enterprise to more consumers to provide services. Therefore, Company A has a greater motivation to find on the right. In addition, a * decreases with t. Intuition is as follows. On the one hand, the power of strategic effects increases; on the other hand, the higher the t, the higher the equilibrium price that consumers start buying from Company A after transferring to Company A. The second effect is less severe in Cournot than in Bertrand (Cournot's competition is less intense, so the new consumer's equilibrium markings are less sensitive to changes in transport cost parameters): In the Bournrand framework, the Cournot framework, Dominant: therefore, the equilibrium distance between the two firms increases with t.

It takes into account a spatial discriminatory one-way Hotelling model in which two firms can move in only one direction. It analyzes the equilibrium that occurs in two different two-stage games: In the game (Bertrand), it is assumed that the firm first selects the position and then sets the price list, where the price may vary from place to place. In another game (Cournot), I assume that the firm first selects the location and then sets the quantity plan, where the number may vary from location to location (Lu &Poddar, 2014).

I show that in both games, a company is located at the end of the market to maximize the number of consumers it monopolizes. However, the equilibrium position of the other firm varies in both models. In Bertrand, a company that is not located at the endpoint is located near the middle of the segment (aggregation never occurs). Conversely, in Cournot, when transportation costs are low, aggregation occurs, and when transport costs are high, a decentralized equilibrium occurs. Cournot has a lower equilibrium distance than Bertrand. Finally, contrary to BHM, I show that the Cournot competition may be superior to the Bertrand competition: when transportation costs are low enough.