Inventory management constitutes a fundamental decision-making problem, especially in systems which must guarantee high availability levels. In complex multi-echelon networks, in case of shortage at a location, resupplying from a near location on the same echelon rather than from the original supplier at the upper echelon, would be a potential faster and then more profitable policy. A wide interest in literature shows the importance of this policy, i.e. the lateral transhipment (LTR), as a means to reduce the inventory costs. This paper deals with unidirectional LTR, which often represent a reasonable policy in scenarios where backorders have different effects on the system. Based on METRIC, this paper defines a system-approach model for determining the stock levels of repairable items in a complex network, by a genetic algorithm optimization process. The model considers non-zero maintenance time for each item and different skills of maintenance centres, in a multi-echelon single-indenture system, with unidirectional LTR allowed. The expert system developed in the paper assists the decision maker to define the inventory levels for the maintenance sites across the system, taking advantage of the LTR to reduce costs and enhance availability. A case study of a European airline shows the relevance of the developed expert system, also considering its reproducibility to face different industrial contexts.