Information revolution provides us with unforeseen opportunities for improving the productivity of services via the optimized planning of production, distribution and delivery. Now companies and clients alike can track and trace mobile resources not only inside their own factories and warehouses but also in all other service facilities and in transit between. Tracking in real time covers all products, vehicles, people and equipment. With ever shortening response times and planning periods, however, the concerns of rescheduling, rerouting, splitting and joining of production batches, product deliveries and value-added service activities will be overwhelming, especially, if realistically counting for the ramifications in time and cost of capacity each activity consumes, including all transfers and set-ups required. To be effective, this kind of time capacitated resource allocation planning also presupposes two properties from production and service resources: mobility and flexibility. In this dissertation, we provide new views and computational methods for the real time planning of production, distribution and service delivery. The new approaches improve capacity utilization simultaneously with more flexible customer service vital for competing in the environment with increasing outsourcing and networking. Efficient capacity utilization, mobility and flexibility are achieved by the simultaneous planning of all required activities and resources by mathematical optimization applied to reliable time-based data. Our approach to capacitated timing balances resource time used for actual production and for capacity consuming set-ups between different production batches or service activities. The explicit consideration of the capacity time consumed by all activities is critical for the realistic planning of high capacity utilization. Mobility of resources in production concerns availability in different time periods, involving costs of setting up and moving back resources through inventory build-up, work-in-process buffers and reserve machines. In service networks, mobility of resources means availability in different locations achieved by moving products, vehicles, containers and service resources, such as cleaning crews or maintenance people, and equipment, among clients, sites and geographical locations. Flexibility of resources is included by allowing production batches or service tasks to be split, joined, rescheduled and reallocated to be performed by any efficient combination of one or more different service resources, such as machines or crews. This dissertation consists of two articles and two essays considering mobile and flexible resource allocation in time-capacitated settings. The first article deals with production planning involving shared resources and the explicit time requirements of the set-ups. The introduction of set-up carry-overs is shown to generate substantial savings in the three key factors of production costs: the number of set-ups, utilization of production capacity and level of inventory. In the second article, vehicle routing problems are solved by minimizing the sum of the traveling cost and the total cost of vehicles actually employed when transportation technology offers scale economies. New methods are introduced for efficiently solving very large problems featuring heterogenous vehicles and time windows of deliveries to as many as 1000 customers, ten times more than in earlier studies. The two essays combine the allocation of shared resources, split tasks and variable set-ups in mobile service operations. The first essay presents a flexible service resource allocation model with a new kind of time-based splitting of work in tasks among available resources. The potential for capacity time savings achievable via this kind of modeling approach is also demonstrated by examples. In the second essay, two different time capacitated resource allocation models for service applications, one with and the other without task splitting, are tested and compared. The tests with a set of synthetic problems indicate up to 33 % savings in the number of identical resources needed when the average length of tasks to split is just over half of resource capacity and the distance between task sites is short. The results imply high capacity savings potential for practical service applications by task splitting. Despite the growing economic importance of time dependent service allocations with mobile and flexible resources, these problems have eluded the traditional modelers due to the technical and conceptual complexities involved. The new modeling and solution approaches suggested here provide some eye opening insights to the general theory while the planning methods with clearly documented results are ready for managerial applications and further development.
|Translated title of the contribution||Capacitated timing of mobile and flexible service resources|
|Publication status||Published - 2010|
|MoE publication type||G5 Doctoral dissertation (article)|
- information systems