Research On Modeling And Applying Of Game Theory On Coopetition Networks Based On Micro-dynamic Mechanism

The competitive pressure enterprises face in current market is increasing, and the behaviors of rivals become diversified and complicated, and at the same time product life cycle becomes short. So how to make optimal strategy in this dynamic and complex environment is an important issue. Enterprises in the course of growth and development need to make appropriate response according to strategies of rivals. Now the competition among companies is based on coopetition networks. So it is necessary to study the behaviors of companies on coopetition networks. Most researches on this issue are qualitative and empirical for a long time, while theoretical and quantitative studies are few. Therefore, main works of this thesis are as follows:In methods and theory:make complex network theory, non-cooperative game theory and evolutionary game theory combined. It is to give nodes of networks subjectivity and to give study of game theory networks structure. So both methods are extended, and coopetition networks can be studied with the methods.In research on coopetition networks:based on the dynamic mechanism of enterprises, study formation mechanism of coopetition networks and build network generation models in order to deepen knowledge about formation and evolution of coopetition networks; based on the dynamic mechanism of enterprises, study strategy response mechanism on coopetition networks under non-profit functions and construct the network dynamic model in order to analyze strategies of companies on competing networks; based on the dynamic mechanism of enterprises, study strategy response mechanism on coopetition networks under profit functions and extend oligopolistic competition model to networks competition model in order to analyze game equilibrium on networks.A variety of research methods are used in this thesis, such as methods of complex network, non-cooperative game theory, evolutionary game theory, Markov chain and simulation. Based on study of this thesis, the main conclusions are as follows:1. If a company is lack of information about cooperation reputation of others, the equal probability selection mechanism is appropriate when it establishes relationship with others; if a company has information about cooperation reputation of others and the cooperation reputation value from 0 to 1 is subject to N (0.5,1) distribution, preferential attachment mechanism of cooperation reputation is appropriate when it establishes relationship with others. Selection mechanism about others is also formation mechanism of the coopetition networks.2. The structure of networks formatted on equal probability selection mechanism is uniform, and its degree distribution follows the Poisson function; the structure of networks formatted on preferential attachment mechanism of cooperation reputation is not uniform, and its degree distribution follows the exponential function.3. If companies have four kinds of action strategies such as competition strategy, cooperation strategy, coopetition strategy and non-contact strategy, when their response functions are non-study function, the limit state probability is equal to the state transition probability at any previous stage; when their response functions are study function, a non-absorbing state will eventually be absorbed by absorbing state with an equal probability of absorption, and the absorbing state limit probability=the absorbing state initial probability+∑the non-absorbing state initial probability* absorbing probability of absorbing state; when their response functions are interferential study function, the limit state probabilities can be obtained by emulator which are different from the study function model.4. In static network game with all capacities used, when all the nodes in the networks reach game equilibrium, sales of the same company in different markets are related to capacities of all the nodes in the networks, and its correlation coefficients vary in value and nature. It reflects the integrity of the network, so small changes in one part may cause major changes in the entire networks.5. In static network game with all capacity used, if the network is fully connected, when all the nodes in the networks reach game equilibrium, capacities of enterprises are divided equally to different markets; if the network is a nearest neighborly network and capacities of different enterprises are equivalent, when all the nodes in the networks reach game equilibrium, sale in every market is shared equally by participants. Because the actual structure of enterprises coopetition networks is generally not completely uniform, shares of different businesses in the same market are naturally different.6. In static network game with all capacity used, if the network is fully connected, when all the nodes in the networks reach game equilibrium, prices in different markets are equal; if the network is a nearest neighborly network and capacities of different enterprises are equivalent, when all the nodes in the networks reach game equilibrium, prices in different markets are equal; otherwise, prices in different markets are different.7. In static network game with capacity constraints, when all the nodes in the networks reach game equilibrium, the equilibrium sale of a market is bigger than its cournot sale, and the equilibrium price is smaller than its cournot price, and the equilibrium profit of a market is smaller than its cournot profits. So it is more competitive in coopetition networks than in only one market where cournot competition exits.8. In static network game with capacity constraints, if the business capacity is bigger or equal than its point capacity and at least there are more than one companies whose capacities are bigger than their point capacities, evolutionary game states are fluctuated and fluctuation is positive with the level at which capacities are beyond the boundary points; if the business capacity is smaller or equal than its point capacity and at least there are more than one companies whose capacities are smaller than their point capacities, evolutionary game equilibrium states are steady and enterprise outputs are equal to their capacities; if there are businesses whose capacities are bigger or smaller than their point capacities, evolutionary game equilibrium states are divided to fluctuated state and steady state.9. In static network game with capacity constraints, if there are businesses whose capacities are bigger or smaller than their point capacities, when evolutionary game equilibrium states are fluctuated, a better response mode is to act against the markets. When the speculator ratio is low, the fluctuation of networks evolutionary game equilibrium states is obvious; otherwise the fluctuation is not obvious. Generally the ratio of speculators in markets is about 0.4-0.6, which is sufficient to suppress the fluctuation of the networks evolution game equilibrium states.