The logistics of preventive health services using fixed and mobile facilities

This dissertation research develops a logistics support system to help health care planners and providers determine when and how to use mobile units to augment traditional fixed-site facilities in delivering preventive services to a regional population. A methodology is proposed with appropriate models and solution techniques. For the locational decisions, we use heuristics from the family of augmentation algorithms where each site is visited randomly and a new solution accepted only if it is better and feasible. Numerical experiments on real data show that these heuristics give good results. These heuristics give better results for the location of fixed facilities than the location of satellite sites visited by mobile facilities. We use the concept of submodularity to explain theoretically the performance variation of these heuristics on different types of location problems. An application of our methodology is given for the Fulton County Health Department (Atlanta, Georgia, USA) which is currently considering a complete re-design of its facility network.