Thursday, July 1, 2010

Algorithmic Management

Algorithmic management!!! So this is what I’ve been doing all along. I’ve spent the last couple of years working on the problem of strategic alignment of services. We’re now looking at the problem of business modeling, and the problem of aligning enterprise architectures. A lot of this work touches on services science, and a lot on questions that have traditionally been in the domain of management (the academic discipline). I’ve always wondered where this sort of work places my research. It’s not traditional computer science or software engineering (although I borrow heavily from these disciplines). It’s not information systems in the traditional sense. It is certainly not management, although that’s the discipline that seems to have raised the questions and identified the research problems. I was telling some members of my group on the train from Brisbane airport into downtown Brisbane a couple of days back – we need a name for this new territory. And now I have it: Algorithmic Management! You heard it here first.

Saving the planet the Green BPM way

I’ve been thinking about “green”. The best way we can address the climate change challenge (outside of some of the “big science” and “big engineering” innovations – e.g., the solar car, seeding the upper atmosphere with various gases and such) is to make better use of what we have. In other words, if we optimize our operations, we save energy – ergo save the planet.

So a possible unit of analysis for us is the business process. These are intended to be descriptions of recipes, or the ways in which we do things. If these could be optimized, we’re in happy territory. For starters, we need to be able to understand the sustainability profile of a given process (e.g., the cumulative carbon footprint). That isn’t easy. How do we accumulate the footprints of the individual tasks in a process model to obtain the cumulative footprint of an entire process? Being able to do that makes all manner of interesting things possible…For instance, I might want to assess the carbon footprint of my process model even as I’m partway through building it. If it looks like the footprint is likely to turn out to be unacceptably high, I might consider designing the process differently.

Part of the problem with assessing the carbon footprint of a process model is the absence of process provisioning information. A process task that seeks to “copy document” might be provisioned in many different ways. I might choose to hand-copy the document. Or I might choose to use a high-speed, high resolution energy-guzzling copier. The carbon footprint is going to be different – depending on the provisioning choices I make. So a solution might be to build a separate resource model, and correlate process steps to this resource model.

That’s sort of stuff I’ve written about in a paper that I’m off to Miami to present (at the IEEE Services Computing Conference). Read about it at