Abstract
In real-world cooperative scenarios, player participation is often partial rather than binary, and payoffs are frequently subject to uncertainty, better represented as ranges or intervals. This paper addresses these complexities by proposing a framework for cooperative fuzzy games with a coalition structure and interval payoffs. The authors derive the Interval Owen Value as a unique solution for fair reward distribution, employing the extension Hukuhara (eH) difference to overcome traditional interval arithmetic limitations. The research demonstrates that for convex games, this value ensures stability and satisfies the Interval Population Monotonic Allocation Function (IPMAF) property, providing a robust tool for decision-making under uncertainty.
Introduction
Classical cooperative game theory typically assumes binary participation—players are either fully in or out of a coalition—and precise payoffs. However, practical applications, such as joint ventures between technology firms, often involve "fuzzy" participation levels where resources are partially committed. Furthermore, the value generated by such collaborations is often uncertain, making interval-valued payoffs a more realistic representation. This paper fills a critical gap in the literature by integrating fuzzy participation, interval payoffs, and pre-existing coalition structures (a priori unions) into a unified mathematical model.
Methods and Mathematical Foundations
The framework relies on several key mathematical innovations to handle the dual challenges of fuzziness and uncertainty:
- Extension Hukuhara (eH) Difference: Standard interval subtraction can lead to invalid intervals. The authors adopt the eH-difference to maintain mathematical tractability and consistency throughout proofs, even allowing for "imaginary" interval numbers to preserve logical integrity.
- Interval Owen Value: Generalizing the classic Owen value, the authors develop a function that accommodates fuzzy membership grades and interval-valued characteristic functions. This allows for the calculation of fair payoffs when players have varying degrees of involvement.
- Axiomatic Foundation: The Interval Owen Value is proven to be the unique function satisfying four fundamental axioms: Carrier, Symmetry, Marginality, and Additivity. This ensures that the distribution of rewards is mathematically "fair" based on established game-theoretic principles.
Key Results and Stability Analysis
The authors validate their model through numerical simulations and illustrative examples. Table 1 demonstrates the input data for a 4-player game, showing how characteristic functions are defined as intervals. Key findings include:
- Uniqueness and Generalization: The proposed formula is the only one that adheres to the established axioms of fairness in this context. Moreover, the model is a true generalization; if fuzzy participation or interval uncertainty is removed, the math collapses back into the standard Owen value.
- Core Membership: For convex games—where the incentive to join a coalition grows with its size—the Interval Owen Value is proven to belong to the "Core." This ensures that no subgroup has an incentive to deviate from the grand coalition, maintaining the stability of the coalition structure.
- IPMAF Property: The value satisfies the Interval Population Monotonic Allocation Function (IPMAF) property. This implies that as the total resources or the game itself expands, individual player payoffs do not decrease, which is crucial for encouraging long-term cooperation.
Conclusion and Future Work
This research provides a comprehensive generalization of the Owen value for fuzzy games with interval uncertainty, offering a powerful tool for decision-makers in complex, uncertain environments. While the model is mathematically sound and ensures stability in convex games, future work could explore alternative indices like the Banzhaf-Owen value. Additionally, extending the framework to handle more complex fuzzy numbers, such as triangular or trapezoidal distributions, could provide an even more nuanced view of probability and uncertainty in cooperative settings.
Summary updated