Content area
Full text
Abstract
This paper introduces a stochastic forecasting model to predict retail sales pre-and-post catastrophic events. The planning horizon begins after observing a tropical depression or disturbance, such as a hurricane declaration, and ends when the hurricane crosses inland area. The true nature of the hurricane effect is unpredictable at the initial stage, but the forecast of hurricane related demand can be updated as with the weather updates. The inventory replenishment can be planned pre-and-post hurricane accordingly. Demand surge during the hurricane may follow a distribution process with an unknown rate. The uncertainty of demand rate can be described by a prior probability distribution based on past data and experts' opinion. The prediction of hurricane related demand surge is challenging. However, forecast may obtain more accuracy if the prediction decision begins at the earlier stage, and if the model continuously revised with latest weather updates according to the Bayesian process. The study suggests many benefits for retail chains, tracking agencies, process industries, private corporations, government agencies; nonprofit organizations that often experience reschedule for services and demand due to post hurricane restoration. A numerical example illustrates the effectiveness of the modeling concept and algorithm design.
Keywords: Catastrophic events, hurricane declaration, Bayesian process.
(ProQuest: ... denotes formulae omitted.)
1. Introduction
This study is motivated by stocking decision problems faced by commercial retail supply chains prior to catastrophic events that are located in the path of a natural disaster. Commonly, retailers react with the customers' demand. However, the demand for the everyday retail products and emergency disaster preparation products suddenly raise after the call of a natural disaster watch, such as hurricanes, earthquakes. Retail products are generally integrated to supply chains, which require ordering lead time and transportation planning for replenishment. If retailers postpone the emergency replenishment orders until the customer demand associated with pre-storm demand surge are observed, they might face the consequence of widespread stockout of emergency supplies and equipment, as well as reputation.
Disaster operations broadly divided in to two different stages; pre-disaster operations like short-notice evacuation, facility location, and stock pre-positioning and post-disaster operations like relief distribution and casualty transportation etc. [1] Because these demands are fairly unpredictable prior to a disaster event, forecast of the inventory replenish quantity becomes complex, which may subject...