|
Abstract
|
In Cloud computing, providers try to sell excess spare capacity of their resources or services to
consumers in a market. An appropriate model raises providers' revenue and rockets customer’s
satisfaction. There are many static pricing mechanism in which providers often considers a constant
price for their services. Whereas providers are willing to use a dynamic pricing mechanism in which
the prices change dynamically based on supply and demand. Therefore, using prediction mechanism in
these economic models with aim to reach a moderate service pricing is vital. In this paper, a new
dynamic pricing model based on Bayesian Vector Auto Regression (BVAR) methods has been proposed.
This model initially forecasts customer’s demand through BVAR method. At the end, it uses a pricing
mechanism to set the current suitable price based on the forecasted demand. Performance results
observed in the simulating and comparing of this model show that the proposed dynamic pricing
schema is able to achieve higher revenue than various other common fixed and variable pricing
mechanisms. Despite of the advantages of the proposed model, the used BVAR method took much
longer to train even though it produced better error results.
|