During the last decade, manufacturing managers have devoted increasing attention to fine-tuning production systems -- both within the plant and across the supply chain. They want to synchronize operations to improve asset management and get the best return on investment. Various techniques -- ranging from visual management systems that simplify in-plant communications to deployment of advanced-planning-and-scheduling (APS) software -- have been used to improve and coordinate plant-floor operations and synchronize activities with downstream demand.
One visual-management tool, widely adopted in JIT/flow production environments, is the use of "kanban squares" -- or kanban containers -- which serve as "pull" signals between one operation and the next. When a square (usually painted on the floor) or a kanban container is filled with work in process, the upstream feeder operation stops producing until space in the kanban area is freed up. An empty square or container is the signal that triggers replenishment by the preceding station -- or, in some cases, from a purchased-materials storage area
For many manufacturers, pull systems have replaced the practice of capacity planning with MRP II systems, which traditionally created work orders for each machine or workstation. Too often, production managers discovered, the forecast-based MRP II-based schedules included unrealistic leadtimes and failed to take into account such plant-floor disruptions as equipment breakdowns or the need to expedite last-minute rush orders. Moreover, the capacity-planning systems typically assumed "infinite" capacity at each station, frequently scheduling unrealistic workloads. As a result, MRP II systems often failed miserably in synchronizing the flow of material and caused large amounts of WIP to build up in the factory.
With the introduction of JIT/flow systems, MRP II became a long-term planning tool rather than a means of scheduling day-to-day activity on the plant floor. For scheduling production in complex -- or multiplant -- operations, the advent of APS systems in the 1990s gave manufacturers a new optimization tool, one that took advantage of more powerful client/server computer systems. Unlike MRP II schedulers, APS systems use finite-capacity logic, basing workloads on more realistic assessments of equipment capabilities. Moreover, the speed of APS systems allows for frequent updates of production schedules to reflect the real-time status of plant operations. And they simultaneously balance multiple constraints, such as equipment capacity and availability of materials and labor.
Some APS planning and optimization systems even take into account such factors as "preferred" customer status in prioritizing schedules, notes John Bermudez, group director for supply-chain research at AMR Research Inc., Boston. "In the past, preferred customer status was rarely considered in the planning cycle."
An APS system "allows a company to harness the advances in computer technology to give it an enormous business advantage," Bermudez says. "Whatever the driver is in a particular industry -- whether it is responsiveness to customers, shorter leadtimes, or better on-time delivery -- APS allows you to offer that. . . . The planning systems we had in the past didn't allow you to merge the business goals of your company with your planning. The factory planner seldom had the business goals in mind. He was just working with 'hot sheets' to expedite [the latest high-priority order]."
The ability of APS systems to facilitate frequent replanning is a major plus. "A lot of problems in manufacturing stem from making the wrong thing at the wrong time -- and having to constantly shift production to handle hot orders," Bermudez notes. "In the short term, manufacturers carry a lot of inventory just because their planners can't react fast enough. If they are doing weekly MRP runs, they have at least a one-week frozen [planning] horizon. And if they are working on a three-week fixed horizon, they need even more inventory to cover unexpected orders. . . . With an APS system, you can replan every two hours -- which allows you to shorten that horizon."
The leading vendors of APS systems include i2 Technologies, Manugistics, Numetrix, Berclain, and Logility. But another 20 or more companies have established a market presence -- and leading firms in the enterprise-resource-planning (ERP) arena have been adding APS capability to their systems, in some cases through strategic alliances or acquisitions. (Baan Co. NV, the Dutch-based ERP vendor, recently acquired Berclain.) By incorporating constraints, as well as business goals, in their optimization algorithms, APS systems typically allow for real-time plan and schedule creation, and also provide real-time decision support and the ability to determine "available-to-promise" capability in a real-time mode.
Different Approaches
APS systems come in a variety of configurations -- some oriented to longer-term planning across multiple locations, others geared to optimizing activities on a short-term basis within a single plant. The middle range of functionality includes supply-chain planning -- on a three- to six-month horizon -- which seeks to optimize the use of manufacturing, distribution, and transportation resources to satisfy both forecasted and actual market demand.
"Our software," explains Nima Bakhtiary, vice president-marketing for Paragon Management Systems Inc., Los Angeles, "does global optimization rather than local optimization. With global optimization, you are synchronizing activities across the enterprise, and the system decides how to execute that. It looks at various constraints on a global level -- including materials, capacity, inventories, and things like provisions in contracts with specific suppliers. You want to be able to look at all the constraints and come up with an answer that meets the corporate objectives."
Systems that emphasize global optimization use "what-if" modeling to answer such questions as where, how, and when to produce products in order to satisfy key business goals, such as improving customer responsiveness or maximizing throughput and return on investment.
"If you are in a capital-intensive industry and you can't easily add new capacity, you can get that into the equation," says AMR's Bermudez. "You can equate your profit level on a particular product to the capacity that a product is using. And you can do rapid 'what-if' analysis to determine whether you should cut back on one product line to push another product instead."
Typically, notes Paragon's Bakhtiary, APS systems draw upon an ERP "backbone" to gather the data needed for modeling and optimization. One of the big issues, he points out, is whether the planning and scheduling system is ERP-centric or APS-centric. With ERP-centric systems, ERP is the primary driver for overall planning and uses an APS engine for local, plant-level optimization. "But with APS-centric planning, the APS software gathers the information, does the optimization, and then tells the other ERP system modules what to do," Bakhtiary points out.
The main difference, he notes, is that APS-centric systems facilitate global optimization, whereas ERP-centric systems take far longer to achieve synchronization between plants -- if, indeed, they have the capability to synchronize. Much of the constraint-management thinking embodied in the new planning and scheduling software is at least loosely derived from the work of author/consultant Eliyahu M. Goldratt , whose analyses are detailed in his widely read books -- The Goal (1986, North River Press) and What Is This Thing Called Theory of Constraints And How Should It Be Implemented? (1990, North River Press) . In simple terms, Goldratt proposed that plant managers identify their primary capacity constraints -- typically the bottleneck operations that limit a factory's throughput -- and then subordinate activities at nonconstraint locations to the pace permitted by the constraint.
The key to improving throughput is to "elevate" the key constraint -- that is, introduce measures to increase its capacity. Once a constraint has been broken, the bottleneck usually shifts to another operation, which then becomes the focus of improvement activity
Goldratt also drew attention to the need to optimize the allocation of work to constraint operations to achieve the product mix with the highest level of profitability.
Maximizing profits
One new optimization software package, which borrows from the Theory of Constraints philosophy, likewise addresses the profitability equation. Rather than focusing on scheduling of production, assembly, or logistics activity, the Maxager System introduced by Maxager Technology Inc., San Rafael, Calif., helps component manufacturers to identify "lost profit opportunity," says Michael Rothschild, president and CEO of the software firm.
"Every company knows its actual profit performance," Rothschild points out. "But what they do not know is what we call the TAP -- total available profit. The gap between actual profits and total available profit [with a given asset base] is the lost profit opportunity."
Using a technique called "bottleneck costing," the Maxager software helps to quantify the lost profits and identify both tactical and strategic options to recapture them. The system collects and analyzes data such as arrival time, setup time, run time, and wait time at strategic control points (the bottlenecks). Then it calculates the rate at which parts flow through the bottleneck operation and the dollar value of those parts by SKU (stockkeeping unit).
"What you really want to know is how many profit dollars per minute are flowing through the constraint," Rothschild asserts. "You want to know the overall profit velocity."
In a factory that makes as many as 1,000 different products, he says, approximately 20% of the SKUs "will be making 500% of the profits. Another 20% of the SKUs will be losing 400% of the profits. And the middle 60% are break-even. And they do not know which is which. They can't figure it out using standard costing."
Most plant managers, he says, have little or no data on the profit velocity at bottleneck operations. "How do you expect to optimize the amount of money that ends up in the company treasurer's office if you don't know the flow rate through a bottleneck?" he asks.
When quality-related delays or machine breakdowns interrupt work at a strategic control point, the Maxager system makes it possible to quantify the cost of those interruptions. "For example, if you could be getting $1,000 an hour through a stamping press, but you have to wait 15 minutes for an inspector to show up, that wait cost the company $250 in lost profits," Rothschild says. "Unless you capture that information you don't know how much money you are losing because you have inadequate quality inspection capacity. And without such information, you can't prioritize improvement projects on an ROI basis." In addition to providing real-time data on the profitability of current operations, he claims that the Maxager system also can help manufacturers determine their potentially most profitable product mix.
