Empowering the pharma planner

Is the Pharmaceuticals sector showing us the way to be truly lean and agile?

Much has been written about lean manufacturing over the years and one of the major wastes that lean manufacturing initiatives target is inventory in all its forms, raw materials, WIP and finished goods. If you have read what has been written, or paid for a lean practitioner to advise you, you will have inevitably concluded that IT will play no part in any lean initiative, and that visual production control (VPC), typically in the form of kanbans and supermarkets, is the only way forward.

We can all learn by looking at how other people solve their problems, and then developing a solution for ourselves based on their experience, but in manufacturing it is rare to get this type of cross pollination between different verticals. For example, how often do staff or consultants move between, say, pharmaceuticals and precision machining? They are typically seen as having no commonality and no one would consider experience of one sector to be useful in the other.

Preactor International is proud of the broad spread of its user base, and unlike most manufacturing professionals we are in a position to compare the different planning, scheduling and production control techniques used in different verticals. One of those sectors is small batch pharmaceuticals production for clinical trials, and using the inventory definition this sector has always been lean, because they have no choice. When you have products and raw materials destined for a single clinical study that may never be repeated, carrying significant inventory has never been an option.

On top of the intrinsic lack of inventory extreme agility is often also required. Say you are running an active pharmaceutical ingredient (API) manufacturing facility. You will have to make simultaneously many one-off small batches of API. Each API manufacture may required a different set of processes and machines, and a laboratory to make it in, all of which has to be cleaned before and after the batch has been made, and will require other resources such as staff who have the required level of training and competencies have to be assigned to the many processes.

Dealing with variables
Quality assurance (QA) checks required at various points during the manufacture cause a large amount of variability in predicted when processes can take place. These checks could obviously mean rework is required, causing a disturbance to the schedule, and in addition the QA resource will only have finite capacity, and thus delays to your process can be caused by waiting for QA resource.

All of this is compounded by the variability in demand caused by factors such as clinical trials which have been planned for months suddenly being cancelled, yet another trial hits the news bulletins and is suddenly top priority, so the production schedule is never repeatable in any way.

Planner power
So if you visit one of these companies will you find the kanbans, etc., which are seen to characterize lean processes? No, the use of VPC is rare in this industry for the simple reason that kanbans are WIP, and VPC essentially creates a make to stock process. Any variation in demand would soon lead to WIP that is not required. So how do they cope with controlling these ultra lean processes and the variable demand? The answer is that they aim for the ultimate lean implementation, make to order (MTO), and in many cases the pharmaceutical sector is bucking the trend for empowering production personnel to make scheduling decisions, and is empowering the planner instead. Is this a good idea?

If you use VPC scheduling decisions are based on empty kanbans. The emptying of a kanban triggers it to be refilled by the operators at the upstream process. This works well with stable demand, however the operators are working in isolation because they have no other visibility of other processes, late arriving orders, etc., nor do they have visibility of company wide key performance indicators (KPIs) such as minimising late deliveries, minimising production costs and so on. Variations in demand will cause problems that the operator cannot be aware of.

Seeing the whole picture
An empowered planner, on the other hand, can see the whole picture and will make scheduling decisions, such as changes in priorities for customers, based on company wide KPIs. With the whole picture the planner can make the trade offs that are bound to be required, e.g., utilisation (fewer changeovers and cleans) vs. delivery performance (small batches), and variations in demand are handled easily. Ultimately the production facility has only to cope with a single KPIschedule adherence, safe in the knowledge that the company KPIs have been taken care of by the planner.

How do we empower the planner? Simple, we give them the tool no lean practioner will mention, a computerised Advanced Planning and Scheduling (APS) system. This will use a finite capacity model, that takes into account both resource and material availability, to rapidly produce a good schedule for the production facility that meets the company KPIs. If necessary the planner can quickly perform multiple what ifs on the schedule, comparing the results to determine the best compromise for the KPIs.

The APS will dynamically aggregate the small batches of the same or similar products that an MTO environment inherently produces, so in our API manufacturing facility we can balance our attempts to minimize cleans with the needs of delivery performance. The aggregation will be re-calculated each time we re-schedule, so we are sure that our process utilisation is as good as it can be under the current conditions.

The planner will re-schedule as often as required by process disruptions and demand changes, often several times per day. In the most planner empowered plants the APS becomes an intrinsic part of the production control system, and no schedule decisions are made without first testing them on the APS. When an issue arises the planning and production staff will gather round the APS to resolve the problem whilst still striving to meet the company KPIs.

Case studies from the pharmaceutical sector show the impact that APS systems are having on what are already lean processes, such as optimised workflow with minimum bottlenecks; massively increased flexibility to quickly re-schedule if order requirements change at short notice, or even mid production run; the ability to run various 'what-if?' scenarios. Whilst not all the techniques used in the pharmaceutical manufacturing sector can be applied to more general manufacturing, we can certainly learn from their use of APS software to make lean processes even leaner.

Dr Nigel Shires is Application Development Manager for the scheduling software solutions company Preactor International (PI). PI, is an independent company based in the UK. First formed in 1984 the company is one of the fastest growing companies specializing in advanced planning and scheduling software and, through its partnerships with other software vendors and system suppliers, offers implementation and software support all over the world.