CVT route optimization algorithm
Telecommunications has had a major impact in all aspects of life in the last years.
There is little doubt that the transformation from the industrial age to the information age has been fundamentally
influenced by advances in telecommunications. Optimization problems are abundant in the telecommunications industry.
Optimization of the routing phone calls is very important in situation of strong competition
between telecom companies. CVT Prepaid call center's goal is to maximize revenue and customer satisfaction,
while minimizing marketing and sales costs. This is a short description of CVT Prepaid Real Time
optimization algorithm.
Set up exceptions base on historical traffic – ASR,
ACD, PDD
Load new carrier prices to database
Find existing routes or create new based on carrier codes
Adjust prices based on historical data QoS
Calculate priority and weight of carrier (Carrier Rank) in route tables based on Least Cost Routing (LCR) or Maximize Revenue Formula, take into account exceptions and cost adjustments (QoS and Cost and other factors)
Load data to switch
Recalculate calling card price
Monitoring real time data, notify operators and customer service about problems, and if needed, automatically adjust priority of carriers in routes base on QoS.
Formula for optimization of revenue per call is very simple: ASR * ACD * (Net Cost Per Min).
ASR * ACD take into account Quality of Service, Net Cost Per Min or Net take into account calling card rates and carrier rates. The whole formula depends from both quality and cost. In case when all carriers in the route have the same quality ASR*ACD = const, we get the same result which we receive from least cost routing.
The whole idea is very close to Google calculation of PageRank for search engine, but in our case we calculate Carrier Rank for route optimization.
The well-known Google idea seems very simple but in reality the real algorithm is very complex and is a well-kept secret. The same situation applies to our phone card business. To provide the optimal routing we should take into account a lot of factors like route type (ground or cell), country, city, historical date time data, real carriers, account service fees etc. So the algorithm of optimization based on the experience, current product and etc is rather art than math.
Another challenge is analyzing real time data, which amounts to millions of transactions per hour, thousands
directions, hundreds carriers and calling cards; this is comparable to the complexity of real-time trading in stock market or space center programs.
By using new VoIP Veraz soft switch, Verso Billing system and our warehouse Vortex software we are able to calculate price, QoS and analyze data in less than ACD time allowing for real-time traffic optimization in the both cost-efficient and quality way.
|
