Use it wisely
Planning & scheduling, MES help manufacturers make the most of machine and labor capacity
By Roberto Michel, editor -- Manufacturing Business Technology, 3/1/2003 12:00:00 AM
The underlying production planning question at Yankee Candle, says Paul Decoteau, has been whether to "chase sales" by building up capacity prior to peak seasons, or smooth out the demand peaks over the course of the year. While the latter strategy carries benefits including consistent labor requirements, says Decoteau, the company lacked the planning tools necessary to follow through with precision.
But that was before the South Deerfield, Mass.-based manufacturer of premium scented candles deployed an advanced planning software application that allows it to model its plant resources and run various "what-if?" scenarios to determine capacity and inventory implications, and generate a constrained master schedule. "The is the first time we've had a tool of this type to manage our capacity," says Decoteau. "We use it to level load—to push down all our peaks."
The deployment of the tool, from Malvern, Pa.-based enterprise suite vendor Agilisys, came at an opportune time. In its 32-year history, the company has grown from a small operation to a rapidly growing company with $380 million in sales in 2001. Yet, says Decoteau, its candle production is still done at its single plant in Massachusetts. While some non-candle products are outsourced, says Decoteau, the planning tool helps Yankee Candle make the most of its assets. "A big plus of the deployment," he says, "is that it helps us control our operations during a time when we've experienced double-digit growth year after year."
Of course, many U.S. manufacturers haven't seen growth in the past two years. Some verticals, such as industrial equipment or telecommunications gear, have faced slow corporate spending on capital equipment. Additionally, even at some rapidly growing companies, management software applications are being used to get more out of existing assets, and to gain insight on where and when to invest in plant equipment.
In this era of doing more with less, capacity utilization gains importance. Yet, say observers, utilization must be balanced against the necessities of order fulfillment. "Planning tools can be used to find optimal lot sizes, or to reduce machine set-up times, but the tools also need to provide visibility into the bottleneck—into what the constraints are," says Carol Ptak, vice president of high-tech and industrial manufacturing for PeopleSoft, a Pleasanton, Calif.-based enterprise suite vendor that offers advanced planning & scheduling tools.
Ptak, who has been a leader in APICS over the years, says the concepts of Dr. Eli Goldratt, first made famous in his book The Goal 19 years ago, convinced many manufacturers that the blind pursuit of "point" resource utilization could be counterproductive, leading to situations like unnecessary inventory buildups in front of constrained resources. Says Ptak, "If you can achieve better utilization on non-constrained resources, that's of little consequence if a bottleneck is holding up operations."
Doing more with less—in this view—requires a balance between resource utilization and bottlenecks. While various applications help achieve this balance, planning & scheduling tools on one hand, and manufacturing execution system (MES) software on the other, combine the optimization and visibility needed for the task.
Capacity questions
Planning tools address capacity issues over different time horizons. A finite scheduling application, for instance, handles short-term sequencing of work. This is the type of factory planning that can reduce changeover times. Over a longer horizon, applications built on advanced planning & scheduling (APS) technology use special algorithms, referred to as "solvers," to address capacity and material constraints, and come up with optimal mid- to long-term plans.
Yankee Candle uses the Agilisys Advanced Planning solution over this longer horizon, says Decoteau, planning in 18 buckets of time (either weeks or months), up to 18 months out. The application's solver, he says, takes a demand scenario and runs it against a model of Yankee Candle plant, equipment, and labor resources. "It time-phases demand, and breaks it down by all our workcenters," he says. "We are much better able to plan ahead, project the inventory we'll need, and assess the impact on our lines."
Yankee Candle follows a best-of-breed strategy. While it uses planning tools and an enterprise system from Agilisys, order entry is handled by a different package, and the company also uses a warehouse management system from Atlanta-based Manhattan Associates. Short-term scheduling is handled via semi-automated means.
While it's possible Yankee Candle might deploy a scheduling solution in the future, says Decoteau, the advanced planning tool has achieved the goal of supporting a level-load strategy. "The tool has helped us realize it's better to level load, and that the cost of the chase strategy—which included having to build up our labor capacity and do a lot of hiring and firing—is higher."
The advanced planning tool's "what-if?" capability, says Decoteau, makes the application not only a capacity planner, but also a "labor and business analysis" tool that allows the company to assess issues like machine investments. But fine-tuning the application's model, he adds, took some consulting help from Agilisys.
After going live with the solution, Decoteau says, Yankee Candle called Agilisys back in for an "audit" of the solver engine. This engagement, he says, lasted about three days, and involved some reconfiguration of the way resources were modeled. "With our aggressive growth, it's a plus that the tool was flexible enough to be reconfigured," he says.
The execution side
Balancing machine resource utilization with bottleneck management at a tactical level, says John Richardson, founder and CEO of CIMNET, a Robesonia, Pa.-based MES vendor, calls for a combination of MES functions and finite capacity scheduling. This is why, he adds, CIMNET just introduced a new scheduling module as part of its Factelligence MES.
"There are many factors involved in managing utilization, but basically you have to be able to schedule the resource, track the progress of work and events, and make adjustments," Richardson says. "The challenge also involves organizing information, and ensuring that the right people are getting the right information at the right time."
Factelligence, says Richardson, offer graphical tools for viewing work-in-process, machine utilization trends, and for managing product definition data. Other factors to consider, says Richardson, are the ability of an MES to roll up and view data from multiple plants, and to have separate production and reporting data stores.
Lou Unkeless, a company vice president for San Jose, Calif.-based Datasweep, agrees that execution applications need to be combined with analytics. In fact, says Unkeless, this realization is why Datasweep, which is known for its Web-centric manufacturing execution capabilities and ability to collect granular "unit-level" data, is ideally positioned to support what Datasweep calls "operations performance management."
There are three components involved in operations performance management, according to Unkeless. One is having the execution and data-gathering capabilities of MES. Second, manufacturers need the right architecture to aggregate, store, and access information. Finally, the analytical tools or a "knowledge dashboard" turn that data into information for operations performance management.
Traditional vendors of business intelligence tools are ill-suited to the role of aggregating and making sense of plant data, Unkeless contends. Operations performance management for manufacturers, he adds, calls for "a top-down strategy with bottom-up deployment."
Costa Mesa, Calif.-based nMetric is another vendor with MES functionality and a Web-based architecture. The company's 4C system combines scheduling tools with monitoring of plant-floor conditions and execution.
Tom Carpenter, nMetric's CEO, says that while 4C has tools for bottleneck management, including a feature called nTrace that analyzes the impact of current conditions of a production order at any level of a bill of material, the strength of the solution stems from being able to constantly monitor and react to events. "4C has an engine that is constantly churning information, but that is cross-checked by a priority engine," says Carpenter. "These priority thresholds are user-definable, because with this type of engine, it's possible to generate too much information."
At Spec-Temp—an Antwerp, Ohio-based manufacturer of tempered glass products that sells to manufacturers of recreational vehicles, manufactured housing, and other markets—deployment of a finite-capacity planning application has helped the company achieve a balance between resource utilization and schedule attainment. According to Brian Miller, systems administrator, the application, from Chippenham, U.K.-based Preactor International, addresses capacity management in at least a couple of key ways.
First, says Miller, the scheduling engine models Spec-Temp's plant resources, including labor, machines, and the time it takes for machines to perform operations. Second, the application's user interface includes a view of the current amount of work, or "loading," on the plant floor, and the progress of work on the plant floor.
Part of this visibility is a scheduling whiteboard function called a Gantt chart. "We're able to easily view shop load, and see which resources are under- or overutilized," says Miller.
But the real power of the application, says Miller, comes from the scheduling engine's ability to model plant resources and apply user-definable rules to the model. "The software knows there may be several different pieces of equipment that can process a part, and knows that some machines can carry out an operation more efficiently than others," says Miller. "However, pure efficiency isn't the final arbiter of what runs where. Other factors, like not delaying orders, can take precedence."
Fine-tuning the Preactor software to embody all those rules, and to contain the proper data on resources, was a challenge, says Miller. In fact, Miller readily admits that this modeling required outside expertise.
Initially, says Miller, Spec-Temp and its enterprise system vendor, Ontario, Canada-based IndustriOS, undertook the task of writing the interface between the enterprise system and Preactor, and implementing the tool. Moving the necessary data between the systems went smoothly enough, says Miller, but neither Spec-Temp nor IndustriOS were yet experts on the tool's modeling.
As a result, says Miller, Spec-Temp called in Preactor's North American master reseller, Quinn & Associates of State College, Pa., to complete the modeling. Several days of this consulting work, says Miller, brought out the ability of the tool to quickly generate optimal schedules. "The Preactor engine can look at almost any constraint you want to put in there," Miller says. "So its strong point is that it's open and flexible, but this also can be a weak point if you're not quite sure how to harness the tool."
Improvements for the record
Since going live with the software, Spec-Temp has seen a 12-percent jump in per employee-hour productivity. Says Miller, "There are multiple factors involved in this improvement, but there is a strong feeling here that much of it came from scheduling improvements.
"Before, we lacked visibility into the impact of events," Miller continues. "If we wanted to take an order, or if a machine broke down, we didn't know what the impact on the schedule would be, and what delivery dates were going to fall off the table. Now we have much more visibility not only on what's happening, but on the appropriate schedules to set."
Finally, says Miller, the deployment has helped Spec-Temp achieve that delicate balance between workcenter efficiency and schedule adherence. Says Miller, "The Preactor engine will pull together resources and materials for as much efficiency as possible without violating due-date constraints. We're able to serve our customers better, and more time is spent on productive activity."
| Who's inside | ||
| Agilisys: www.agilisys.biz | CIMNET: www.cimnetinc.com | Datasweep: www.datasweep.com |
| IndustriOS: www.industrios.com | nMetric: www.nmetric.com | PeopleSoft: www.peoplesoft.com |
| Preactor International: www.preactor.com | ||
