Joint optimization of maintenance policy and two-way stock transshipments policy for balanced systems

Traditional optimization models of maintenance and inventory policies only focused on a one-way stock transshipment policy (i.e., longitudinal transshipment) of common systems, ignoring the effect of lateral transshipment on maintenance policy. This paper formulated an integrated optimization model of maintenance policy, longitudinal and lateral transshipments for balanced systems. The failure mechanism of balanced systems is different from common systems in the interdependency between the unit and the corresponding unit on the symmetric position. Once a unit fails, the symmetric unit must stop working to keep it balanced. To make full use of the failure characteristics of the balance system, an effective rearrangement policy is proposed to improve the system's reliability, and replacement actions are conducted to timely correct failed units caused by random environment shocks. To timely replenish the inventory of spare parts, not only a longitudinal order policy (s₁,S) from the depot to any base but also a lateral order policy (Q,s₂) among different bases are simultaneously considered. Since the order times, replacement time and rearrangement time are random variables following general distributions, a semi-Markov decision process framework is utilized to formulate a minimum cost model by selecting appropriate order policies. A modified value-iteration algorithm is proposed to solve the integrated optimization model. Finally, a comparison analysis between with and without the lateral order policy is employed to illustrate the priority of the proposed policy by a simple Unmanned Aerial Vehicle system.