Investigation of supply disruption with time dependent parameters

Ying Rong
2005-01-01
Abstract:Although supply disruption sounds very rare, it happens all the time. Besides those factors, such as weather, labor strike, terrorism, when the manufacturer production scale is very large, such as shampoo produced by P&G, those small retailer may not get their order at all due to randomness of manufacture and their small scale. Song and Zipkin [2] models this situation as discrete queueing system. Since those small retailer cannot get information of manufacture process whether it occurs disruption or manufacturer randomness, we can approximate the whole system as a disruption model. The most research done so far assumes constant disruption rate and recover rate. Under constant disruption rate assumption, people like Parlar and Perry [1] studies how it affects those classic inventory policies, such as (r, Q) policy. Synder [3] provides a tight approximation for the lost sales EOQ model with disruption (EOQD). Song and Zipkin [2] study the case where supply system can be modelled as a discrete time Markov chain. After reducing state space, they shows the value of knowing information of supply system. Recently Tomlin and Synder study the supply system with multiple fail rates following discrete time Markov chain. They shows the state dependent order up level will gain significant saving again constant order up level policy.In this paper, we consider the availability of one single supplier following continuous time Markov chain and one retailer facing Poisson demand. The events, such as adverse weather, strike, machine breakdown, congestion of orders from all retailers the supplier encounters, can cause the supplier incapable to provide its …
What problem does this paper attempt to address?