Abstract: The reservation data of the car rental system cannot represent the real demand of customers. Inventory control leads to the mismatch between supply and demand, and also produces the "overflow" and "reappearance" of demand. Unconstrained estimation usually takes short-sighted or strategic assumptions, which results in insufficient consideration of bounded rational behavior of customers. Taking the car rental demand of multi-price class as the research object, decision is formed based on the bounded rationality of customers. The customer choice behavior is analyzed by Logit model, and the customer utility is obtained by combining the value function of prospect theory, so as to analyze the demand transfer behavior and realize unconstrained estimation. The calculation example shows that the multi-price class spill improved model can effectively avoid demand overestimation, and explain the impact of real demand on revenue. The stock control revenue is increased by 0.81%.

Key words: transportation economy, unconstrained estimation, Spill model, car rental, bounded rationality