Is a Multiagent Orchestration Platform for Robotics in Your Future?
Guest blog by Dan Gilmore, CMO, Roboteon (MHI Member)
With the growing interest in and adoption of robotics in distribution and manufacturing has come this important realization: you need a different type of software to manage, orchestrate and optimize processes across robots of different types and vendors.
A few years ago, the analysts at Gartner recognized that trend, and put a name to what they considered a new category of software, which they called a “Multiagent Orchestration Platform,” or MAOP.
What is an MAOP?
Gartner acknowledges it is an evolving concept but has defined MAOP like this: “Multiagent Orchestration Platforms act like intelligent middleware that integrates and orchestrates work between various business applications, heterogeneous fleets of operational robots, and other automated agents like doors or elevators. These solutions orchestrate and assign work and monitor and coordinate the activities of diverse fleets of robots.”
Gartner adds that “These solutions will assign work to the right robots based on the characteristics of immediate and prioritized tasks and communicate with other types of automation (agents) like door or elevator controls.”
The MAOP is an add-on to, not a replacement of, your Warehouse Management System (WMS), focused on robot integration and optimization, and therefore productivity, utilization and throughput.
The development of the MAOP term and concept was led by well-known Gartner analyst Dwight Klappich, who retired in late June.
In a recent Warehouse Automation podcast from robotics management software provider Roboteon, Klappich said that there were four key building blocks or components of MAOP solutions, which are as follows: integration, orchestration, optimization, and observation.
Integration, Klappich said, “is just the ability to easily connect various different types of agents between business applications, robots, and other types. So that for example with automated truck unloading, somehow I need to get a message to the truck unloading robot to start the work.”
The second building block is orchestration, Klappich said, which involves the fact is that “once I can integrate, I need to be able to orchestrate that work because the reason people are introducing robots and things like that is they want to get work done, right? The robots are there to support the execution needs within the warehouse or plant.”
He added that “what I’m doing one day could be very different the next day. That’s one of the main reasons people are buying mobile robots instead of conventional automation. So the heart of the system is that orchestration piece, that ability to define those end-to-end work processes and how things need to connect, not technically connect, but from a process standpoint, between those steps in the process and be able to seamlessly execute that.”
The third component, optimization, comes “because you realize, well, if we can just execute the work, the next step is we want to do it the best way,” Klappich noted.
He added: “We want to take those processes and understand all the work that might need to be done and start to apply much more robust decision support or simulation logic to those processes to say I’m not just doing them, but I’m doing them in the most efficient and productive way possible.”
Finally, on the fourth building block, “observe,” Klappich said that “if you look at most of the robot companies, they have their own software stack and most of them have some pretty decent analytics capabilities, but the problem is it’s only for their robot platform.”
He added “what I am hearing is that companies want a single pane of glass where they can kind of look across all the activities in their warehouses and then start to apply analytics at that level.”
Good stuff from Klappich, who also acts as a Strategic Advisor to Roboteon.
There’s a lot more on this on the full podcast.
And yes, there is a very good and growing chance an MAOP will be in your future.
