FURP – Master Production Plan
This prototype, developed in AIMMS in the begining of 2002, supports the automatic generation of integrated and optimized master production plans for pharmaceutical drug lines at FURP – Fundação para o Remédio Popular (freely translated as Popular Pharmaceutical Drug Foundation).
The pharmaceutical drugs are produced based on “active pharmaceutical ingredients” (API), substances with pharmacological effects. The API’s used by FURP are acquired according to several different types of contractual arrangements, each one with specific regulations establishing, for example, maximum delivery deadline and the minimum and maximum commercial transactions involved. Generally, the best prices are achieved through the use of contracts that allow for a longer delivery time such as national and international auctions. On the other hand, the need for resupplying raw material stocks on short notice can demand the use of direct purchases, which are significantly commercially restricted.
Production is distributed throughout the available lines according to the types of drugs: tablets, coated pills, capsules, creams and ointments, suspensions, oral rehydration solutions, injectables, granulates, eye drops, liquids (glass and plastic) and sunblocks.
The pharmaceutical drug production must meet strict hygienic and sanitary conditions. Changes from one product to another (and even between lots of the same drug), for example, require that the production lines be cleaned, according to previously established procedures, consequently reducing productivity. To minimize machine setup costs, production is conducted in minimal lots (or batches). On the other hand, lots must be correctly dimensioned according to demand and product shelf-life. Due to demand seasonality, part of FURP’s production is outsourced, with final product quality guarantees.
FURP is the largest government pharmaceutical drug manufacturer in Brazil; in 2006 its production reached 2.5 billion pharmaceutical units and 3,2 thousand municipality’s, with over 5 thousand clients. Amongst them, State Health Departments – in particular the state of São Paulo – state and municipal hospitals; municipal consortia; municipalities; municipal, state, federal and philanthropic institutions (and in particular the Ministry of Health); and unions and foundations.
Moreover, FURP is responsible for supplying pharmaceutical drugs for the State of São Paulo, through the “Dose Certa” program, a Basic Pharmaceutical Assistance Program for the Department of Health of the State of São Paulo. Since 1995, Dose Certa has distributed, free of cost, over 11 billion pharmaceutical units, essential for basic care units, such as vitamins, creams, analgesics, antipyretics, antibiotics, anti-inflammatory and drugs used to treat hypertension diseases.
The master plan generation process at FURP seeks to roughly determine the production volumes of each finished product (SKU) for the subsequent month to the current one. Until 2001, this process was fundamentally based on the use of spreadsheets – at the time, no computational Supply Chain Planning tools (or Advanced Planning tools) were available for this type of activity.The manual generation of a Master Plan is a complex and laborious task, given the production allocation flexibilities and operational constraints that must be considered, such as:
- the plan should maximize demand meeting goals through its various distribution channels;
- the production plan should be coherent with the BOM (Bill of materials) and the raw material availability, which is proportional to available and quarantined stock and supply scheduling;
- the receiving schedule should consider current contract deadlines and values; while the medium and long term supply, must be compatible with the API acquisition characteristics;
- the plan must be compatible with the finite production line capacities (including preventive maintenance schedules);
- the plan must meet lot sizes and preparation time for each product;
- the plan must consider outsourcing possibilities.
Since products compete for scarce shared materials and resources, these conditions can only be guaranteed through the use of optimized policies, such as, (a) reduction in setup time; (b) effective use of materials; (c) eventual production anticipation, observing final product and API shelf-life times and production line availability variation and (d) rational use of outsourcing. In case this isn’t enough to totally meet demands, distribution channel priorities are established, such as the “Dose Certa” case.
Due to this methodological gap, UniSoma proposed the development of an Operational Prototype that (1) illustrates to FURP the possibilities of representing the master production plan problem through the use of mathematical programing and (2) indicated the potential gains through the use of Supply Chain Planning tools (or Advanced Planning tools) by the PPC Department.
The prototype was developed in AIMMS and contained three basic layers: (1) a mixed integer programming model to generate the integrated production plan at FURP; (2) a mathematical solver (CPLEX© of ILOG Technology) to automatically solve and optimize the scenarios and (3) a graphical user interface (GUI) through which the planner could interact with the mathematical model, changing input data and visualizing generated results (what-if analysis).The main results from the prototype, simultaneously and coordinately generated, are:
- the monthly master production plan for each SKU;
- the level of demand met for each month and distribution channel;
- internal and external monthly SKU production;
- SKU stock at the end of each month belonging to the planning horizon – typically 12 months;
- buffer/safety stock (in days) for “Dose Certa” demand;
- production lines monthly use – considering the setup times due to the production schedule;
- monthly supply plan, for each critical material and each contract type – to guarantee plan generation agility, only API’s were considered critical material; raw material such as packaging have been disregarded during the model validation studies, although this can be achieved by simply registering them in the BOM;
- monthly critical material consumption plan;
- critical material stock for the end of each month;
In principle, the model was conceived to guide the solver in the search for solutions that would maximize the service level. This criteria showed itself insufficient in the selection of more advantageous purchasing contracts (balancing prices and deadlines) as well as investigating “cheaper” outsourcing possibilities. Due to this, an economical objective was adopted, using the NPV (net present value) of the global marginal contribution at FURP during the planning horizon. This alternative allowed the generation of a solution in which non-intuitive trade-offs can be observed. Besides this, various analysis were made available (“what-if” analysis), such as:
- Evaluation of the setup time impact on the global marginal contribution of the company;
- Safety stock cost studies;
- Quantification of the impact on financial costs;
- Optimized maintenance allocation activities.
To compute the NPV of the global marginal contribution, several economical parameters were made available, such as, for example, capital cost, raw material purchase costs through the use of different types of contracts, external outsourcing costs and final product prices.
The development of the prototype, in a 4 month deadline, overcame its established objectives. Through it was made possible to demonstrate to FURP the advantages and importance of adopting optimizing techniques to ensure that the whole of the business requirements be met.
|AIMMS Service Partner UniSoma|