Intelligent Manufacturing € June € 1996 € Vol. 2 € No. 6

James River Revamps Scheduling with Intelligent Supply Chain System

• improve customer service to the point of delivering perfect service;
  
  
• reduce finished goods and work-in-process inventory to drive down the need for working capital;
  
  
• shorten planning cycles to produce faster and better production schedules at the company's 24 production plants;
  
  
• and increase asset utilization by generating more stable schedules to optimize the capacity of existing equipment.

"Our system was generating unreliable schedules mainly because it was using a batch mode approach," explained Andy Charchenko, James River's manager of product supply. "These were totally inadequate for meeting changes in today's marketplace."

The system was report-driven rather than using on-line scheduling techniques. Planners at James River were scheduling from computer-generated reports rather than directly from the computer screen. Instead of highlighting exceptions for action, these reports required manual intervention. So if a plant produced 500 SKUs (stock-keeping units), planners and schedulers had to wade through 500 sheets of paper to find the problems.

The system was spreadsheet-dependent. Moreover, the planning and scheduling systems at various facilities were not integrated. Methodology and practices were not consistent across all businesses. The system's ultimate anachronism was that it was transaction-based.

"The system generated a list of reports, so planners would highlight errors, correct them and rerun the reports," said Charchenko. "These changes would result in further errors that had to be corrected. It was an iterative process. But we were not getting any closer to real solutions."

James River's goal was to replace the outdated system with a decision-support environment. They chose a system developed by Numetrix Ltd. (Toronto, Ont., Canada), a provider of intelligent supply chain management software solutions, because its software modules provided useful interfaces with James River's existing systems and linked up other non-Numetrix products as well. In addition, Numetrix's Supply Chain Integrator (SCI) module could integrate current and future systems that would optimize production and distribution scheduling across the entire enterprise.

"We wanted an open-architecture solution -- the ability to interface with our legacy and other systems," recalled Charchenko. "When you are trying to bring together data from various sources to make business decisions, creating effective interfaces with the different software systems becomes extremely important."

Numetrix consultants installed three Numetrix Software modules at James River. Linx, a global strategic and tactical optimizer module, was installed as the monthly planning tool for both the towel/tissue and the Dixie line of products. Planx, the master production planning module of the Numetrix Solution, was chosen as the weekly scheduling tool for James River's Dixie product line.
Finally, Numetrix consultants installed the SCI module to serve as the interface between both the Linx and Planx modules and existing James River systems, in addition to managing data and preparing reports.

With the modules in place, each solution came with a cost penalty based on finite constraints. The financial impact of these decisions will help to optimize the whole network by giving James River the total delivered cost of the product.

"Before, our existing system gave us only the relevant manufacturing costs or distribution costs, not both of them together," said Charchenko. "Now, we can increase our profitability because we know the costs -- variable manufacturing, distribution, storage and handling -- attached to each solution."

Output from James River's old system was comparable to rough-cut production schedules created by earlier MRP (manufacturing resource planning) software that produced an educated guess of a company's production capacity. At that time, there was just not enough computer power to put in all the necessary constraints. But the Planx module includes finite constraints that generate realistic production schedules. Each solution comes with real costs attached.

After installation, James River wants to add more "horsepower" to the three RISC workstations on which the new system operates so the hardware can crunch the numbers even faster. Currently it takes about 30 minutes for the system to create a monthly production schedule. Charchenko would like to see weekly schedules produced much faster than that. "I know it will never be instantaneous because of all the data that must be digested," he said.

It is still too early for James River to quantify the benefits of the new system. "We don't have any hard measures as of yet," said Charchenko. "But the early feedback we have received tell us that we are on track to achieving our goals. Thanks to dynamic sourcing, we have already found ways of reducing our overall manufacturing costs. For example, we have two facilities that make the same line of products. By focusing directly on the manufacturing and distribution costs at each plant, we discovered that one of them had the capacity to make all of the product at a lower cost."