CONCLUSION
This article demonstrates how optimization and simulation
techniques can be adopted in retail store sales prediction
and micro-competition analysis. Based on GIS platform,
we develop a more elaborate analysis framework than traditional
gravity model. Simulation technique provides the
means of directly evaluating the impact of these factors.
Besides the sales prediction of new stores, this framework
can also provide an explicit estimate of the impact of new
stores on existing competing stores, which is of more value
for new site evaluation work of business development people.
Real case has proved the effectiveness of the framework.
In our framework, we deal with the stores in the same
way and ignore their differences. However, multiple factors,
including visibility, store size, service level, environment,
etc. can influence customers’ behavior greatly.
Therefore, how to build a quantitative model to evaluate
the stores’ influence on the customers will be an interesting
work.
In current framework, we perform the sales prediction
on given candidate sites of new stores. In real case, how to
select those candidate sites is also a complex analysis work.
Therefore, another interesting extension of the framework
would be incorporating sales prediction into automated site
location.