On the performance of GRASP and dynamic programming for the blood assignment problem

Blood is very precious and vital in medicine; blood transfusion occurs every day during surgery or for other reasons. However, the continuous daily demands for blood transfusion in the hospitals have exerted more pressure on blood banks. For this reason the assignment of blood at blood transfusion centers requires optimization techniques in order to have efficient allocation as to reduce wastage and meet all the demands of the patient. The constraints of acquiring blood (cost), the blood grouping into types (Rhesus factors) and the limited shelf life of blood have pose a big challenge to blood transfusion centers. This paper models blood assignment problem in form of multiple knapsack problem and implement two local search optimization techniques, dynamic programming and GRASP, that use the model to facilitate efficient assignment of red blood units from available blood types to different request for blood of different types. Therefore, the result shows that GRASP handles data more efficiently over dynamic programming.