摘要
The purpose of this paper is to generalize Landscape theory proposed by R. Axelrod and, then, to apply it to the civil aviation industry for simulating alliance formations in it. Landscape theory provides a well-known agent-based simulation model for analyzing alliance (or coalition) formation process. When a set N of agents or autonomous decision makers is given, the theory assumes that each agent tries to make a coalition in such a way that the resulting alliance minimizes its frustration. The theory is essentially based on two premises. One is that a propensity is symmetric, i.e., the propensity of agent i toward j is exactly the same as that of j toward i for any agents i and j in N. The other is that the number of alliances is restricted to two, i.e., at any moment N is partitioned into two parties. Though the two basic premises underpin the theory and make the model simple and operational, they do not always reflect alliance formation processes in a realistic way. A generalized Landscape theory that this paper proposes removes them and allows asymmetric propensity and existence of alliances of any number. Since the premises are essential for the model, the generalization requires a drastic reconstruction of the whole idea of the theory. Finally, we analyze a real alliance formation process in the civil aviation industry. This analysis provides interesting insights about the industry as well as some validation of our generalized Landscape theory.
The purpose of this paper is to generalize Landscape theory proposed by R. Axelrod and, then, to apply it to the civil aviation industry for simulating alliance formations in it. Landscape theory provides a well-known agent-based simulation model for analyzing alliance (or coalition) formation process. When a set N of agents or autonomous decision makers is given, the theory assumes that each agent tries to make a coalition in such a way that the resulting alliance minimizes its frustration. The theory is essentially based on two premises. One is that a propensity is symmetric, i.e., the propensity of agent i toward j is exactly the same as that of j toward i for any agents i and j in N. The other is that the number of alliances is restricted to two, i.e., at any moment N is partitioned into two parties. Though the two basic premises underpin the theory and make the model simple and operational, they do not always reflect alliance formation processes in a realistic way. A generalized Landscape theory that this paper proposes removes them and allows asymmetric propensity and existence of alliances of any number. Since the premises are essential for the model, the generalization requires a drastic reconstruction of the whole idea of the theory. Finally, we analyze a real alliance formation process in the civil aviation industry. This analysis provides interesting insights about the industry as well as some validation of our generalized Landscape theory.
