Heuristic Solutions for Locating Health Resources

Abstract

Of the three metaheuristic strategies tested that can help determine healthcare facility locations, scatter search performed the best and fastest. Hypoglycemia, or low blood sugar, can cause many health problems, including blurred vision, mental confusion, and speech impairment. If someone becomes severely hypoglycemic and doesn't recover quickly, he or she might lose consciousness and go into a diabetic coma. If that happens, the person must get treated within 15 to 20 minutes; otherwise, he or she almost certainly will suffer devastating physical damage - possibly neuropathy, or blindness - or might even die. Because the risk of permanent neurological deficit increases as the coma is prolonged, it's important for people with diabetes to live no further than 20 minutes (known as the critical time) from their closest health center. Facility location problems such as this involve determining where to install resources and how to assign potential customers to those resources. Most studies on location problems are framed under deterministic conditions. Our proposed solution is more realistic. We adapted, implemented, and compared three metaheuristic strategies - scatter search, tabu search. and variable neighborhood search - to find the best locations in Spain's Burgos province to place health resources for treating people in diabetic comas. To check the efficiency of SS, TS, and VNS, we used instances of the well-known OR-Library as well as real data from the Burgos area in northern Spain. Using metaheuristics is a good option when the problem's complexity prevents us from using a commercially available solver to solve it exactly. This is especially true here, because we're considering hundreds of locations.