5. Discussion on existing data:
To ensure realistic results, we sourced parameters values from the regional blood bank in Abu Dhabi. Costs incurred for collection, handling, and storage of units are estimated to be around 800 AED for a 450 ml blood packet. A weekly demand of 700 blood units that include all blood types is regularly provided to hospitals affiliated with the regional blood bank. A minimum inventory of 100 unit of each blood type is kept for emergencies. The most recent data indicates that a total demand of 708 blood units with variant shares for each blood type: 220 blood units for O-, 53 for O+, 98 for A-, 98 for A+, 88 for B-, 88 for B+, 53 for AB+, and 10 for AB-.
The blood type O- is a universal donor; therefore, extra quantities are always kept by blood banks for that specific blood type since it can cover shortages for other blood types.
The data acquired from Abu Dhabi regional blood bank also stated the minimum inventory levels for each blood type as the following: 6 blood units for O+, 11 for A- and A+ each, 10 for B- and B+ each, 1 for AB- since there are fewer people with AB- blood type, and 6 for AB+.We considered the feed to the blood bank is stable from regular donors.
We considered the feed to the blood bank is stable from regular donors. According to the blood bank, the frequency of people donating blood is different for each day in the week, with Sundays registering significantly higher donation levels. Wednesdays are the lowest in donations while stable donation levels are registered for Mondays, Tuesdays and Thursdays. To quantify that, we have set some percentages for each day, which the blood bank approved, and these are: 33% of blood units are collected on Sunday, 10% on Monday, Tuesday, and Thursday, 30% on Wednesday, 5% on Friday, and 2% on Saturday.
The main purpose of this model is to optimise the number of daily collected units to meet the scheduled demand and avoid waste. As per the blood bank in Abu Dhabi, keeping a waste level of 1% was among their KPIs.
Code in Gurobi (Anaconda)
For our project, and as part of the course requirement, we have used Gurobi to optimize our problem. This being said, and since we are dealing with 8 excel sheets (1 per blood type), we had to call the following libraries: