Customer Information Driven After Sales Service ... - RePub
Customer Information Driven After Sales Service ... - RePub
Customer Information Driven After Sales Service ... - RePub
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
4.5. Numerical Experiments<br />
plexity is significantly reduced due to the local search. In Section 4.5.3, we discuss the<br />
computational efficiency aspects of the MDP approach, RMH+, and RMH.<br />
4.5 Numerical Experiments<br />
In this section, we analyze the performance of our proposed RMH against the MDP<br />
approach, RMH+ and existing Spiral Router heuristic (SRH) at IBM (described in Section<br />
2.7.1). According to SRH, each incoming demand should be fulfilled from nearest<br />
non-empty stock location. From stock location perspective, SRH heuristic is of FIFO or<br />
FCFS nature where each incoming customer, regardless of the contract type it posses,<br />
gets a spare part. From customer perspective, the objective is find nearest neighbor<br />
stock location that is non-empty. We should note that SRH is conceptually similar to<br />
nearest neighbor heuristic proposed by Bertsimas and Van Ryzin (1991). The authors<br />
propose nearest neighbor heuristics for service engineer dispatching problem, where each<br />
incoming demand request is to be served by nearest available service engineer. Although,<br />
the formal proof of optimality is not provided, the authors show that nearest neighbor<br />
heuristic performs better than other execution heuristic for service engineer dispatching<br />
problem. There are two major differences between service engineering dispatching<br />
problem considered by Bertsimas and Van Ryzin (1991) and current problem settings<br />
of spare parts logistics execution. First, service engineer resources are reused to service<br />
future demand requests. In other words, a service engineer who is being considered for<br />
the incoming job may be sent from his current idle location or his current working location<br />
(i.e. customer location). Due to this, the current service engineer dispatching<br />
decision is dependent on preceding dispatching decision for given service engineer. In<br />
spare parts logistics, a part is always sent from either of the non-empty stock locations<br />
(fixed locations). Secondly, the analysis performed by Bertsimas and Van Ryzin (1991)<br />
does not account for customer base heterogeneity.<br />
As mentioned in Section 4.1, our goal in this chapter is to develop a technique that<br />
accounts for various aspects of customer heterogeneities in spare parts logistics execution.<br />
The heterogeneities in the customer base exist due to distinct contractual agreements and<br />
customer locations. Due to the distant service contracts for a single spare part type, the<br />
customers pay different service prices for the maintenance service. Similarly, the penalty<br />
costs associated with service deadline violations are also different per service contract.<br />
109