The Ultimate Scheduler

Manufacturing scheduling is set to get a big boost with the help of a real-time computational organizer.

		NRC Business Case Challenge 2006
		Transforming Technology into Licensing Opportunities
		Saving their Skin
		Harnessing Bacteria to Battle Pollution
		The Ultimate Scheduler
		Mathematics and Biology Unite to Save Lives and Reduce Health Care Costs

Think scheduling your busy personal life is a difficult juggling act? Try efficiently scheduling a diverse manufacturing plant. It's estimated that just trying to calculate all of the possible alternate schedules for 24 manufacturing jobs on one machine would take your personal computer about 20-billion years.*

A team of NRC researchers has created what could be a revolutionary computational approach to manufacturing scheduling. The team is on track to commercialize the world's first distributed, real-time manufacturing scheduling technology.

The NRC manufacturing scheduling technology recently checked-off a milestone in its own daytimer by participating in the licensing category at the 2006 NRC Business Case Challenge.

"There's a major gap in the software industry to address real-time manufacturing scheduling. It's a crucial information technology issue for manufacturing companies competing in the global market," says Dr. Weiming Shen, the NRC Integrated Manufacturing Technologies Institute (NRC-IMTI) team's senior research officer. "Our technology will enable manufacturers to quickly respond to changes and disturbances which cannot be predicted in advanceon the shop floor, and thus maximize their productivity by optimizing their equipment utilization rates."

What sets the NRC scheduling software apart is that it addresses dynamic changes and disturbances in a timely manner, without disrupting the overall operations. All existing manufacturing scheduling software programs on the market are based on a predictive, centralized scheduler. This type of system assumes a specific global set of conditions, such as that all of the equipment will be working — something that in practise, is rarely the case.

In contrast, Dr. Shen's London, Ontario-based NRC-IMTI group have developed dynamic, real-time manufacturing scheduling software based on intelligent software agents.

With an agent-based manufacturing scheduling system, each machine, operator, robot, or other manufacturing component is represented by its own software, or agent. The agent software communicates and cooperates through a wired or wireless network with as many other agents as necessary to troubleshoot scheduling problems. Thus, as much as possible, decision-making is decentralised. If, for example, an assembly line robot breaks down, the nearby local hub of agents can determine the quickest and most feasible solution without having to disrupt the entire production.

"Rather than trying to calculate the absolute overall optimal solution, the agent-based system uses an economics-inspired approach to quickly calculate a best-case solution and thus keep production running and avoid idle machines or employees," explains Shen, who with 14-years of agent-based manufacturing R&D experience is a global leader in the field.

Global scheduling is carried out collectively by the agents through a negotiation mechanism and protocol. Managers and operators monitor and adjust the scheduler as necessary via a user-friendly interface.

A key strength of the new technology is that it's scalable, says Dr. Shen. It can be applied to a single shop floor, to a company's entire distributed operations, or even across companies. It can also be programmed to handle manufacturing components at various stages from the supply chain to final delivery.

"From the software point of view it's the same idea, the same system architecture and the same technology but implemented at different levels," says Dr. Shen, also an adjunct professor at the University of Western Ontario where he co-supervises an international lab of about ten graduate students.

He says the scheduling software's scalability will be essential in making the product cost-effective for users.

"Our product's cost won't be like buying a single scheduling software package with a one-size-fits-all approach," says Dr. Shen. "The cost will be relative to the number of individual agents involved. So a smaller manufacturer with fewer agents will pay less for the product than a much larger manufacturer."

Dr. Shen and his team mate Dr. Qi Hao are targeting the world's 4000 tooling shops, with several hundred in southern Ontario alone, as the first market for the agent-based manufacturing scheduling software. This highly-competitive manufacturing sector produces just-in-time components for automotive and other sectors, making machine scheduling a major factor in cost competitiveness.

The NRC scheduling technology is currently a demonstration version. Dr. Shen and NRC-IMTI Business Development Officer Ms. Daisy Fung are now negotiating with several companies to choose the site for the first real-world testing. They're also looking for a Canadian software company to commercialize the technology.

From previous experience, Dr. Shen is aware that integrating the agent-based scheduling software into a manufacturing process – and the human adaptation this entails – will likely require more work than it took to develop the program. But, he says, global competition is pushing companies to achieve greater and greater efficiency, and agent-based manufacturing scheduling software will be a key tool to achieve this.

"I'm confident in this technology," says Dr. Shen. "Within ten years, in order to compete globally, all manufacturers will have to be using this type of software."

Recommended Links :

• Doing Business with NRC
• NRC's Areas of Research: Manufacturing
• NRC Integrated Manufacturing Technologies Institute
  

		How would you rate this article?
		12345 1 = not informative5 = very informative Comments: