From the above, it appears that both Maximum capacity constraints have shadow prices (Pi) of 0; therefore, they are considered nonbinding constraints. Likewise, the inventory equilibrium and minimum inventory level constraints have non-zero shadow prices and 0 slack; and are therefore considered binding constraints. Actually, the slack value of 0 for the inventory equilibrium is evident since it’s an equality.
7. Conclusion and Future work
This model proved to be able to solve the inventory management problem for a perishable product despite its simplistic approach. The algorithm could be further expanded to include other types of blood products.The sensitivity analysis showed which parameters hold the largest effect on the blood bank performance and what are the intervals in which the blood bank could play to keep its collection plan optimal. This project could further be improved by implementing an inventory update that would help to track the inventory and regularly check whether the collected units were consumed totally or not. This leads to another constraint in which the perished units are kept to a minimum so that the blood bank meets its KPIs. A notion towards integer programming model would prove to be also a possible future enhancement for the current model.
Acknowledgement :
We would like to deeply thank Dr. Andrei for his support throughout this project, from its initial formulation to the implementation of the code and its results. Warm thanks go to Abu Dhabi blood bank as well, mainly Dr. Naima Oumeziane for her stimulating collaboration, and her tireless answers to our endless questions and concerns regarding data and results.
References
1.Adewumi, Aderemi, et al. “Optimizing the Assignment of Blood in a Blood Banking System: Some Initial Results.” 2012 IEEE Congress on Evolutionary Computation, 2012, doi:10.1109/cec.2012.6256633.