2009/04/25 by Michael O. Ball, Maurice Queyranne · 6 citations
Computer Science · Business, Management and Accounting · Decision Sciences · #Optimization and Search Problems #Supply Chain and Inventory Management #Advanced Bandit Algorithms Research
paper · doi:10.1287/opre.1080.0654
In this paper, we consider the revenue management problem from the perspective of online algorithms. This approach eliminates the need for both demand forecasts and a risk-neutrality assumption. The competitive ratio of a policy relative to a given input sequence is the ratio of the policy's performance to the offline optimal. Under the online algorithm approach, revenue management policies are evaluated based on the highest competitive ratio they can guarantee. We are able to define lower bounds on the best-possible performance and describe policies that achieve these lower bounds. We address the two-fare problem in greatest detail, but also treat the general multifare problem and the bid-price control problem.