Finite scheduling works – well

Remember when MRPII (manufacturing resource planning) fell out of favour? Suddenly every vendor worth its salt was selling ERP (enterprise resource planning) instead even if, to the disinterested observer, the systems looked well nigh indistinguishable. The same ‘rebranding’ exercise is clearly swinging into action today with finite capacity scheduling (FCS) and advanced planning and scheduling (APS). As a result technical distinctions are being lost under a general assumption that APS means ‘leading edge, fast and superior’ whereas FCS is a synonym for yesterday’s technology. Typical is the comment of Ralph Seeley at industry watcher Cambashi: “FCS is the last resort of a company whose processes are tortuous, ill-considered and inflexible. It is an IT Band-Aid applied to a festering wound.” Ouch! This judgement does, however, underplay two salient points. Firstly, after years of struggling with infinite capacity planning, any of the newer scheduling technologies – regardless of the name – comes like a visitation from a guardian angel. Secondly, many people actually using the ‘FCS Band-Aid’ say that it has brought an unprecedented level of order and productivity to their shop floor processes. The counter-argument runs that while they may seem to be doing a good job, because most FCS systems work in conjunction with MRP/ERP, they are merely compensating for the inherent idiocies of the core system, rather than curing them. FCS exponents ask where’s the problem? MRP solves material problems and manages the overall plan, while FCS gets on with producing the detailed, ‘do-able’ schedule that supports it. Increasingly good two-way interfaces mean there is little delay or problem in reconciling the two. This is a split of responsibility that is fiercely opposed by some of the newer ERP vendors. They put the scheduling engine at the very centre of their software’s operation, with a simultaneous handle on materials, production resources and labour. To them it is completely illogical to pin the workings of your business on a planning infrastructure that needs the contents of two separate ‘brains’ to solve a single problem, even if they are swapping information at a fairly high level. It’s a bit of an over-simplification, but these two separate approaches to scheduling probably make more useful a start point than the uninformative APS/FCS label. So let’s take a look instead at the advances being made in each of the two ‘camps’. In doing this, I am deliberately focusing on the kinds of solutions that are within easy reach of SME manufacturers. There are some powerfully sophisticated and productive planning systems available today which optimise entire supply chains but the cost simply rules them out for the majority of companies within this sector. First of all, the ‘bolt-on’ approach. Many ERP vendors have chosen to go to the scheduling specialists. Some, like Symix, actually chose to buy their scheduling provider – in its case Pritsker. Others have supplied niche developments either as rebadged products or through partnerships and third-party sales. One of the foremost players in this market is Preactor whose scheduling tools work in conjunction with most of the major mid-range ERP systems. Its ability to produce capable scheduling solutions at extremely keen prices has seemingly left little incentive for many ERP vendors to develop their own. Indeed, it has recently introduced an advanced scheduling package, Preactor APS, that handles multiple constraints and material control for around the price of a good estate car! It harnesses complex rules to minimise WIP (work in progress) and job lateness with preferred sequencing to minimise changeover times and deal with wandering bottlenecks. It also has some features to help overcome the issues of dealing with multiple ERP/MRP systems. For example, it can take into account the availability of materials at any stage of the process, automatically linking orders together based on how a product is made and pegging material usage. Nonetheless, Preactor is an unashamed add-on. Anyone buying it will live with the issues of data interchange and if ‘real-time’ is your obsession, it is probably not for you. As, however, our box-item shows, there are companies of good standing who like eating this particular pudding! Some established vendors of Preactor have, however, recently taken a different tack. One of them is enterprise software vendor Fourth Shift. Its move is part of a wider game-plan, taking the best-of-breed route to supplying a web-enabled SCM (supply chain management) product suite. Using its existing ERP software as the backbone of the system, it has added a web-based configurator, and CRM (customer relationship management), field service and HR (human resources) packages, all from third-parties. It is also offering an APS product from Systems Modelling Corporation’s Tempo technology. Though it has changed suppliers, Fourth Shift still maintains that specialist scheduling wins out over embedded technology any time. Sales director Simon Philpott says that the company has moved because the new licence agreements give it control over all ‘relevant product development’ and allow it to provide single-source support. Philpott maintains this route not only gives good proven technology but also overcomes the commonly perceived disadvantages of ‘bolt-on’ technology – the need for an interface and therefore a lack of ‘real-time’ activity. He says that dedicated development between the two vendors allows data from the two systems to be synchronised almost immediately. “You can do it for every ‘what-if’ simulation or even every minute, if that’s what you want to do.” He says there does come a point where the user actually has to commit to a schedule before launching it back into the ERP system but even a fully integrated approach requires the user at some point to change a simulation into a firm plan. SME ERP software developer Swan has also left the Preactor stable but for different reasons. Its recent acquisition by a German company (the combined UK organisation is now known as infor:swan) has produced some marked changes in product direction. Parent Infor has a marked preference to full integration rather than best of breed, particularly in planning and scheduling. The planning engine in its infor:com ERP product takes simultaneous account of available parts, machinery and labour, forward or backward scheduling from a resource list that combines bills of material and work-centre routings. In effect, this means that bottlenecks are likely to be flagged up sooner and can be ‘unblocked’ by a backwards run to start competing jobs earlier. In a marked departure from the classic MRP process, Infor:swan’s Doug Miles says that the system can be run against individual sales orders rather than combined materials requirements, allowing much more flexibility in handling new or urgent jobs. The newer brand of ERP vendors goes even further in decrying the non-integrated approach. Maz Mohamedali of Lilly Software is scathing: “You simply can’t bolt on scheduling because it’s at the heart of what you do.” Lilly actually holds the patent on concurrent scheduling of materials and resources and its development, Visual Manufacturing, treats the APS planning engine as a core component of its ERP software. Works orders, material requirement planning and resource allocation are all driven by the schedule. “Companies tend to believe that although their business is running well, they have a problem with sorting out their shopfloor. So, many of them regard a scheduling tool as a stand-alone way of treating a departmental ailment by ordering jobs to make a sensible flow. What they miss is the chance to take it one stage further by using it to make real business decisions. Like ‘which job is worth more money to us’ or ‘which of these possible schedules is the one that will give us the most customer satisfaction?” explains Mohamedali. “For example, most schedulers say they consider materials and resources together, but what happens when you get a problem loading a job at a specific machine? With our integrated approach, I can instantly see in my scheduling window that the problem is not machine availability but lack of material at the time I will be loading it. I can click on the job to find out what it needed and drill down to the specific purchase order for that material. “I can chase the supplier or quickly run a couple of alternative scenarios to assess the impact of freeing up material by pushing back another job. I can even drill upwards and see which of the two customers makes us more revenue before deciding on the best schedule. In other words, I can get the whole picture. If you travel the conventional MRP/FCS route, you will probably have to go away and re-run MRP to see what material is coming in overall before you can even start reworking the schedule.” Mohamedali admits that working this way makes different management demands. “The master scheduler will play with the software and come up with several different scenarios. Deciding which schedule is going to run may only take a few minutes out of each day but it becomes a business decision that could involve purchasing, sales and finance. It needs to be taken by someone who sees it from the perspective of its impact on the whole company rather than the capacity constraint of the shop floor. It’s not a new need created by technology, it’s a level of decision making that has probably always been there but has been obscured by the conventional ways of operating.”

• Boots gets better with Preactor Boots Contract Manufacturing (BCM), the manufacturing arm of the High Street chemist, uses Preactor at three of its factory sites and plans to expand to two more. BCM produces thousands of different products for Boots and other retailers. With some 100 manufacturing vessels in the Airdrie factory alone, production scheduling is complex and demanding. Preactor has been interfaced with BCM’s existing systems (Hoskyns MAS/MCS/ MRP system and QAD’s Mfg Pro). Work orders are downloaded from the MRP system on a daily basis. The planners work with Preactor to sequence the orders according to due date, either backwards or forwards, by using standard stock holding or Just-In-Time methodology and taking into account available machine capacity. As part of this process, an ‘earliest start date’ can be stipulated to ensure that over-prompt production does not increase finished goods inventory. “One of Preactor’s several strengths is the way it takes account of plant availability,” says Nicky Cox, advanced systems developer at BCM. If a line or machine breaks down, planners can quickly re-run the schedule to try to reconcile due date and reduced capacity. Preactor also takes account of labour availability, allowing casual labour and overtime shift working to be calculated according to workload requirements. Orders can be sequenced according to their ‘physical attributes’ (for example, white fill before blue) using matrices set up within the scheduler to minimise changeover. According to Cox, by scheduling for maximum productivity, Preactor contributes greatly to better customer service and improved profitability.
• Wired for visual manufacturing Cambridgeshire based JDR Cable Systems makes sub-sea cables, hoses and umbilicals for the offshore energy industries. Its range includes communication and life-support links between divers and the surface. It is a highly diverse, make-to-order environment. Three years ago, JDR replaced a bespoke manufacturing control system and scheduler – described as ‘a glorified spread sheet that didn’t link into material on the purchasing side’ – with Lilly’s Visual Manufacturing. Technical director Patrick Kearney describes the difference it has made. “Visual schedules all our machines, operators and materials. It won’t let you start a job unless there is capacity and materials so it keeps inventory and work in progress at a very low level. If a material isn’t in stock, the scheduler lets you take the default time for normal delivery from supplier and schedule it, but at the same time it generates a list for the buyer to purchase the material. “We can re-run the schedule in under 10 minutes. Before we would probably have had a production scheduler working for a day and a half to do one ‘what-if’. It made us reluctant to do it too often – now we do it several times a day. It helps us find better ways of handling normal and unexpected workload. We can see immediately how a late order will affect not only the schedule but also our turnover for the month. It means we can easily assess the options open to us in, say, contacting other customers to ask if a delay is acceptable, or seeing if the work justifies the overtime we will need to get back on track. “Every product is a one-off but we weren’t learning from our previous work. With Visual, you build an engineering master with all the material, machines and set up and run times for the product. Once you have built it, you can go back and compare your estimate with your actual times and costs. It saves so much time.”
• FCS versus APS Finite Capacity Scheduling (FCS) tends to be used with MRPII. MRPII creates a master production schedule (MPS) of products to manufacture, and the plan is checked for feasibility against the capacity of a few key resources – rough cut capacity planning. Next you explode the bills of materials from the MPS to establish the schedules for manufacturing and raw material receipt. Using FCS, planners can take these plans and, with detailed information about manufacturing at each work centre, material availability and the rest, find feasible and even optimal schedules. FCSs provide graphical tools to let planners ‘see’ and move production orders between work centres, and find the best plan as situations change. But this is the point: strictly speaking, FCS focuses only on the capacity constraints, making no allowance for material availability as a constratint. In FCS this has to be determined separately after the production schedule. APS (advanced planning and scheduling) is different inasmuch as it tries to find optimal solutions instantly by solving for a variety of constraints simultaneously: classically materials and capacity, including specific human resources. It harnesses mathematical techniques like Linear Programming and requires today’s modern computers so that all data is memory-resident for fast computation.