Agent technology to optimisation engines

Supply chain planning and optimisation systems could be about to get a great deal smarter. Newcomer Magenta is launching ‘multi-agent’ software technology that it claims has already proven its ability to strip more cost and waste out of supply chains, particularly in logistics. Agent technology, which started to hit public awareness in the late ‘90s, is growing rapidly in application, according to Prof Mike Wooldridge, head of Computer Science at Liverpool University. “Software agents do delegated tasks and co-operate and work as teams in real time, always negotiating with other agents to get the best deal,” he explains. And the significance of that in supply chain terms is that it can deal at a micro level, where existing systems are stuck closer to the macro level. “Most supply chain systems use solver algorithms as the engines to provide planning and optimisation,” insists Magenta CEO Jon Himoff, who jointed the company 18 months ago to push the technology. “They’re good,” he says, “but their limited, and that’s why manual intervention and workarounds comes in because people have been better at sorting out some of the dynamics.” Himoff reckons the potential for agent technology to deliver improvement there – with real-time, responsive systems and additional functionality saving substantial cost and time – is immense. And hence Magenta’s system, launched this year. It takes agent technology and encapsulates it in Java and J2EE for enterprise deployment, the objective being to provide ‘smart virtual marketplace’ technology able to deal with resources, like trucks but arguably also production capacity and people, and automatically negotiate and renegotiate optimised ‘deals’ right up to the wire. Consulting firm Credo is already using the software, but for business process modelling. It’s harnessing Magenta’s Network Designer capability to solve complex business problems which couldn’t be solved by conventional modelling. In this case, the system links existing data on orders, assets and resources with an accurate representation of business rules. By flexing those, Credo can understand the impact of different strategies operationally and from a financial viewpoint. Credo will use Network Designer in a number of client situations. These include helping to deliver true end-to-end optimisation of complex supply chains, designing flexible and cost-efficient distribution networks and aligning pricing strategy with operational capacity. The software will also help model complex, outcome-based outsourcing bids to understand the risks, constraints and financial impact of different scenarios. Meanwhile, Himoff expects the technology to be transformational for logistics and supply chain management. “Logistics behaviour, sales strategies and so on are all configurable,” he says. “And the software allows operators to run any combination of scenarios to develop contingency plans against possible changes to delivery schedules.” Magenta is offering systems to end users, but is focusing its attention on the software community in much the same manner as Ilog and Preactor, expecting the technology to be embedded.