Interop Case Study #1: “Fitness Landscapes and the Electrical Grid”
As previously noted, “Interop” is part of a series of experiments regarding the format of a book. In the case of Interop, we publish it along with a set of case studies that have served as the raw data for the analysis and theory we present in the book version. As our early case studies on digital music, digital identity systems, and mash ups, all the materials are freely available online, via SSRN. Over the next few weeks, we will introduce some of the more recent case studies, several of them authored by wonderful research assistants on our Interop research team. Today – in time with the very high temperatures outside and the increased energy consumption – we would like to introduce the case study on the “smart grid” written by Paul Kominers. Paul submitted the following abstract to introduce the case he researched with us:
Imagine a mountain climber without a map. Rather than going downwards to return to town as might be sensible, he wants to find the highest point in the entire mountain range. But due to a blizzard, he cannot see very far; he can only tell whether a certain direction takes him higher or lower, and he has to stop every fifty yards or so to reevaluate and pick a new direction.
We can imagine many things happening to this mountain climber. He might find a path to the very highest point on the range. But he easily might not. If he finds himself near the top of the second-highest peak, he will probably follow the path to the top of that peak. In fact, if he finds himself approaching any peak, he will probably follow the path up to the top of the peak, unless the blizzard clears and he can see that he is on a lower peak. If he can see where the higher peak is, he might go down far enough to get onto a path that takes him up the higher peak, and then go straight up. Alternately, he might be lazy. He might find himself somewhere that gives him a vantage point to see where the highest peak is, but decide that he does not want to expend the effort to go from his current vantage point to the highest peak. Good enough is good enough for him.
This is an analogy for a fitness landscape. A fitness landscape is a way of picturing an optimization problem. A problem is analogized to a set of coordinates, with each coordinates having a certain fitness value; this builds up an imaginary, n-dimensional landscape. The goal, then, is to find the fittest point on the landscape, but limited sight range and unwillingness to invest to make big changes can get the search for maximum fitness stuck on a suboptimal or local peak.
Our electrical grid is much like the mountain climber stuck on his suboptimal peak. If we had had the sight range to see the next generation of technology back when we built the grid, we would have taken a different path to build that technology in the first place. And big changes require costly investments of time, money, and political capital, all of which are scarce. Further, the path is not perfectly clear. Although we have every reason to believe that a revised electrical grid would pay off substantially, exactly what a revised grid should look like is an incredibly hard problem. This makes moving even harder.
The next generation of the grid is the Smart Grid, a grid built up of intelligent, interoperable components. The Smart Grid comes from making individual grid components more self-aware of what they are doing at every moment. They can then be joined in networks that allow them to make better decisions about how each of them uses and transmits energy.
Currently, a great deal of discussion and debate is taking place. Government regulators, members of industry, and other stakeholders are coming together to discuss what a new, interoperable grid should look like. They are working together to design a more efficient and responsive electrical grid. In essence, they are finding the path to the higher peak.
Please check out the full case study on SSRN.
