- eg. By 31
st
March of 2018, monthly demand will be made available for
the years FY 2018-19, 2019-20 and 2020-21. Similarly, by 31
st
March
of every year, monthly demands for the next five years will be made
available.
b) Approach to be followed for the expected results:
The assignment will entail development of statistical models for all the
major States of the country. The modelling exercise is expected to
produce the following:
i) Total Expected Demand of the State to be met.
ii) Expected supply from Renewables & other must run sources.
iii) Net power, the state is expected to source from competitive sources
(Diff. of i) and ii) ).
4. Proposal invited from interested parties clearly mentioning the following:
a) Capabilities of the party for such exercise; past experience etc.
The vendor will submit information regarding any such assignment or similar
assignment involving development of such models, undertaken by it
recently.
b) Methodology & tools used.
- The methodology to be followed and various tools to be used by the
vendors will be informed, such as:
- Description of the model/ combination of models to be used
- Capability of self-learning feature of the model.
- Use of machine learning/ artificial intelligence based models
- Requirement of data: data requirement for the exercise will be defined
clearly by the vendor including the following:
Historical data requirement
Type and structure of data
Periodicity of the data
Source of data.
- The vendor will also mention whether the data is available with them or
the same will be sourced from public source or procured from authorised
sources/ statutory authorities.
- Vendor is expected to take care of the data and other requirements and
provide the outputs in the form of various reports as described above.
c) Expected accuracy
- The vendor will mention the expected accuracy of the output.
- Further requirements for improvement of the accuracy will be clearly
spelt out, which may include:
o Additional data requirement
o Estimated time period, particularly with models having self-
learning features.
o Any other requirement