|
| |
How we approach a problem...
Our modelling philosophy is to use as much relevant information as possible.
Since load is determined as a consequence of many factors and many decisions by
many people, we believe that accuracy in forecasting requires that we take into
account a large number of relevant factors.
To manage information on a large scale while keeping out models very responsive,
we have both specialized modelling methods and software, and a unique way of
viewing aggregate load that helps us to bring multiple tools to bear on the
problem.
We view load as arising from multiple processes, which can be grouped into four
categories:
Identifiable load is the load arising from identifiable, quantifiable
influences, mainly consisting of human behaviour within a physical environment
that can be predicted based on the clock and calendar, weather effects, and
effects arising from macroeconomic and demographic considerations. The
appropriate techniques for forecasting this component of load are very large
scale, highly-parameterized regression analysis, time-varying parameter
estimation, and Bayesian estimation.
Latent load is the load arising from slow moving processes that can be readily
observed in the data, particularly once the identifiable load is removed, but
which cannot reliably be attributed directly to any particular factor. The
techniques we use for this component are Box-Jenkins analysis with time-varying
autocorrelation parameters, path analysis, principal components analysis,
response surface estimation and other latent variable techniques.
Exceptional load is the load arising from exceptional events, where the events
can be predicted, or are known in advance, and there is a body of experience
from similar events. Appropriate techniques for this component of load include
neural nets, scenario analysis, and intensive close client contact and
reporting.
Unpredictable load is the load arising from unforeseeable events, or from
foreseeable events for which there is no body of experience. Useful techniques
for this component include fuzzy logic systems, contingency analyses, and Monte
Carlo simulation.
TESLA is more accurate than any other load model of which we are aware. We
believe that no other load model can be as accurate as TESLA without employing
the same range of tools as TESLA.
Very Large Linearized Systems (VLLS)
To handle the complex interaction of many nonlinear relationships, we have
developed a process we refer to as very large linearized systems. The VLLS
approach allows us to build a model that is responsive, information-intensive,
and conformal to complex, irregular response functions.
Weather Effects
TESLA treats weather effects in particular with great care. We have developed
methods to capture and analyze weather effects, and to incorporate them in our
forecasting models over multiple horizons.
Data Handling
No model can perform better than the data it depends on. Our experience combing
through the data to find and fix the errors and omission that inevitably arise
is an essential part of accurate work.
Platform Tailoring
A good tool should be easy to use, and we tailor our software for each user
site. Our systems can function on PCs, UNIX workstations, or mainframes.
|
