Electrical Generation Unit Commitment Planning

The electrical unit commitment problem is the problem of deciding which electricity
generation units should be running in each period so as to satisfy a predictably varying demand for electricity. The problem is interesting because in a typical electrical system there are a variety of units available for generating electricity, and each has its own characteristics. At one extreme, a nuclear power unit can provide electricity at a very low incremental cost for each additional megawatthour of energy, but it has both a high cost of starting up again once it has been shut down and it takes awhile to bring it back up to full power. A typical nuclear unit may be shut down only in the Spring or Autumn, when there is very little heating or air-conditioning demand, so demand is lowest. At the other
extreme, a gas turbine generator can be started up in a few minutes. However, its
incremental cost per megawatthour is much more expensive.
The obvious policy is that as demand increases, we first turn on the efficient, but costly to start generators and lastly turn on the least efficient, but cheap to start. As demand decreases, we shut down units in the reverse order. Decisions are a little more interesting if there is a modest peak in demand of short duration. Then it may be economic to skip an intermediate unit and instead turn on an inefficient, but cheap-to-start unit for the duration of the short spike. Various other features of various types of units also complicate the decision. We partition generator units into four types in order of the complexity of their features.