| Almost all published work on revenue management deals with forecasting and optimization methodologies in the absence of competition. For a taxonomy of revenue management problems, see Weatherford and Bodily [120]. Kimes [561 and recently, McGill and van Ryzin [72] give an overview and survey on revenue management.; Research on competitive aspects in revenue management is sparse. Netessine and Shumsky [77] pose a revenue management game with two players and two costumer classes in the airline context. In the economics literature, Borenstein and Rose [19] and Dana [31] study price dispersion and advance-purchase discounts in the U.S. airline market.; The first part of this thesis addresses competition. We present and critique a simplified version of the traditional hotel revenue management problem. We extend this model to include demand substitution and market competition. We show that there exists an equilibrium which can be attained via an auction mechanism. Furthermore, the perfect competition rate structure is bounded by the optimal monopolist rate structure. Results are obtained by exploiting linear and integer programming theory, duality theory, dynamic programming theory, and by use of numerical experiments.; Competitors to hotels need not own any hotel rooms. Even if a company is practicing revenue management, we demonstrate conditions under which third parties, such as tour agents, can profitably buy room reservations from the hotel and sell them to travellers. When a hotel can detect the middle agent, there are times when the hotel should let the agent participate and conditions when it is advantageous for the hotel not to let the agent participate. Here, we follow a Stackelberg game theoretical approach (Binmore [12]) when formulating the optimization models and again use numerical experiments.; The second part of this thesis addresses revenue optimization in the absence of competition for situations not covered in the literature. We prove the integrality property of the linear programming formulation when demand is known and the problem is to find optimal room allocations that maximize revenue. We propose a fast and efficient network flow algorithm as a solution technique. Standard results from integer programming and network flow theory are used to achieve these results.; We then turn to include no-show behavior when dealing with overbooking issues in a multiple-night-stay environment. We show that in certain scenarios, extending the common bid-price approach to consider multiple night stays and available capacity leads to small increases in expected revenue over that achieved by constant overbooking limit heuristics. We use simulations and, for one of the proposed bid-price approaches, a simulated annealing approach to find optimized parameters. |
