The core of the program is the algorithm based on the theory of reinforcement learning.
The Reinforcement Learning concept is an area of machine learning in computer science. It centers on what actions need to be taken in a specific environment to maximize cumulative reward. Reinforcement learning is particularly well suited to problems, which include a long-term versus short-term reward trade-off. It has been applied successfully to various problems but is unprecedented in the Hospitality Industry, and Revenue Management in particular.
Thus, the core algorithm of the system is based on maximizing the end result, i.e. achieving the highest possible yields taking into account all revenue-generating departments and the hotel’s financial structure.
i-Rates is not designed to forecast the exact number of expected potential customers. Instead, it constantly evaluates which of the typical demand levels is expected for every day in the future. This is based on a wide range of constantly changing dynamic and static parameters.
The system has 3 levels of adaptation to the market conditions:
The basic level is a library of typical demand scenarios (developed for each property and constantly updated)
Our model utilizes sophisticated computer algorithms to reveal booking patterns, demand level scenarios, and market conditions specific for each property. This is achieved by in-depth computer analysis of the hotel's booking history. Based on this historic data, the system "restores" property's demand levels, variations of elasticity, and builds a library of typical scenarios, which are strategically applied to each day of the year, 365 days ahead.
Second level: switching between scenarios
In the second level we go further, making sure that applied demand scenarios correspond to the real sales levels. If the system finds that a current sales pattern is a better fit for a different scenario, it will switch and act further to maximize revenue and profits based on the new scenario.
The system's flexibility allows the manager to change the scenario if he or she obtains external information about demand levels that were not accounted for previously.
The third level is a constant strategic monitoring of the current number of remaining rooms to sell, and reactive actions (instant response to changes in capacity).
This module evaluates a wide range of statistical characteristics. It monitors customer flow at any given moment and responds immediately to any reservation, cancellation or other change in capacity. It also takes into account competitive pricing. The system’s flexibility allows the manager to fine-tune these evaluations based on additional external information.
This module provides the maximum interaction between the manager’s professional experience, and the computers functions. The program compares the professional’s experience and additional market information, evaluating multiple metrics to reach the highest possible outcome.
These 2 levels of adaptation are comparable to the adaptation process of live intelligence. Building an artificial intelligence based on the same principles is the core of our system. This provides stability and adaptability. The final product of these adaptation levels is dynamic recalculation of the most optimal rate for each day in the future. Based on each property's financial structure, the program maximizes not just revenue but the property's final profit.