KAVA asked Solvice to develop a planning optimization algorithm to distribute the on-call shifts for 4930 Belgian pharmacies fairly across Belgium over a half-year period. Solvice implemented a combination of intelligent planning moves and swaps, along with efficient meta-heuristics to produce optimal plans in minutes.
The Belgian territory is divided into 126 sectors. However, a pharmacy can belong to multiple sectors. In every sector, one pharmacy should have on-call duty. The distance between the selected pharmacies on duty should be minimal. Pharmacies have desiderata indicating whether they want to be on call. Candidates for the night on-call duties are those who have to be on-call during the day.
After a thorough analysis of the requirements and available data, Solvice conducted interviews and workshops to clearly describe the balance among the different objective functions to all actors involved.
"Integrating OnShift has really helped us to manage our staff more efficiently and has increased our ability to make the best use of our staff."
Solvice implemented a combination of intelligent planning moves and swaps, along with efficient meta-heuristics to produce optimal plans in minutes. The assignment of on-call duties to all pharmacies for half a year effectively incorporates all objective functions. In addition, the custom solver is parameterized to allow Kava to easily change the balance/weights between different objective functions in the future.