A Service Supply Chain (SSC) may be described as a network of service provider facilities (in-house or outsourced), each of which is able to process one or more service tasks on an as needed basis. Two key characteristics of a SSC are (i) the business service is decomposable into several sequential tasks that can be processed by different service providers, and (ii) the primary capacity resource is skilled labor. SSCs are increasingly being developed by companies that experience a high variability of demand for their services (e.g., loan processing, analytical consulting services, emergency repair crews, claims processing, etc.). Typically, the customer wait time penalty is very high, to the extent that if the service is not provided within a certain time, the customer service request will abort. As a result, the service provider needs to maintain sufficient processing capacity to meet peak levels of demand. The primary advantage of a SSC, relative to a traditional dedicated facility, is that the processing capacity (labor) can be economically adjusted (lower hiring and firing costs) to match changes in the current demand level.
In this dissertation, a hierarchical framework for modeling the decision structure in SSCs is developed. This framework introduces and defines the key SSC entities: service products, service jobs, service providers, and the parameters for characterizing the demand behavior. As part of the framework two problems are formulated and solved. First, given that Service Supply Chains are intended to be dynamic delivery networks that efficiently respond to demand variations, a strategic problem is which candidate service providers are selected to form the SSC network, and how the service tasks are assigned within the provider network. The problem is formulated and solved as a binary program. Second, a consequent tactical problem is how the workforce level at each service provider is dynamically adjusted (hiring and firing) as the real time demand data comes in the problem is formulated and solved as a linear program that bounds a mixed integer program (MIP).
The strategic model takes the demand parameters, the competing providers' information, and the service and tasks parameters, to select the providers that are going to become part of the SSC and assign tasks to them. A method to quantify cumulative demand variation per seasonal cycle is presented to derive aggregate demand parameters from the forecast. The design objective of the strategic model is to minimize set up cost and projected operational cost. The objective is achieved by simultaneously minimizing capital cost, hiring cost, firing cost, service delay cost, excess capacity cost, labor cost, and quality cost while fulfilling the capacity, tasks assignment, facility installation, and task capability constraints.
The tactical model is constrained by the providers and task assignment resulting from the strategic model. It uses a more accurate demand forecast, and minimizes actual operational costs represented by hiring cost, firing cost, backlog cost and labor cost, while fulfilling the production balance, routing, capacity, workforce balance and demand constraints. It is solved in two phases. A relaxed model is solved as an LP and its solution is used for bounding a MIP problem.
Finally, the behavior of the two models is studied by performing numerical experiments changing key supply chain parameters such as hiring and firing cost, demand variability, labor cost, and backlog cost.
|
![[NJIT logo]](http://archives.njit.edu/vhlib/images/njit_logo.gif){kind=logo}
![[Van Houten Library logo]](images/vhl-top-image.jpg)
