Supply chain activities problems naturally consist of multiple decision
modeling agents, which are interrelated in a ‘‘hierarchical’’ structure in which
individual activities are often governed by separate supply chain entities. Thus,
they often have conflicting objectives. Such a hierarchical decision network
structure can be mathematically illustrated by using ‘‘multi-level programming
(MLP)’’ principles, in which a decentralized decision system is studied.In this
paper, the behavior of a decentralized three echelons supply chain consisting of
multi suppliers, one manufacturer and multi heterogeneous, independent retailers,
who are involved in procurement, production and selling one type of product in
separate and independent markets is studied. With a leader-follower scenario, and
no direct connection between the retailers and suppliers, the net profit of the chain
entities is maximiz. This notion is initially modeled as a multi-level non-linear
multi objective Stackelberg game. Since this model is NP-hard problem,
Stackelberg solution is obtained by using a hybrid algorithm consisting of two
phases. In the first phase, multi-level non-linear multi-objective problem are
solved via a fuzzyMax–Min decision model for generating Pareto optimal
solution. In the second phase, a search algorithm is developed to find the best
values of lot-size and number of replenishment cycle for supply chain entities
based on the result which has obtained in the first phase. By this method which is
proposed in this paper, the inventory, manufacturing and pricing policy are
simultaneously determined. Moreover, sensitivity analysis of the results is
performed by applying a real case study problem.