چكيده به لاتين
This thesis developed competitive supply chain netwotk design problems under uncertaninty. Studied supply chains are single product and single period with forward (three and four layers) and closed-loop (six layers under government intervention) flow of product under Hotelling and Huff customer utility functions and simultaneous and Stackelberg games. The supply chains are considered by different structutes namely: centralized, decentralized and cooperative modes and are modelled by non-convex mixed-integer nonlinear programing.
With respect to the fact that the models are composed of different players we use multi-level programming to formulate the problems and as they are newcomers to the market,they encounter a high degree of uncertainty and lack historical data, so we use fuzzy mathematical programming to cope with the uncertain parameters. Moreover we proposed different algorithms constructed based on the Lemke and Howson algorithm , Wilson algorithm, variational inequality formulation, bi-level programming, nested bi-level programming , Nash concept, Stackelberg concept, the modified projection method, and the possibility theory. We derive the equilibrium condition and establish the finite dimensional variational inequality formulation, and provide properties of the equilibrium patterns in terms of the results of existence and uniqueness. Finally, we generate some instances inspired by a real world study to discuss the effect of the different structures of the competitors, namely centralized, decentralized, cooperative, or unknown modes, on the equilibrium solution.
