Abstract
We consider the following multi-unit selling problem: several identical units of some commodity need to be sold, one at a time, to a stream of arriving bidders. Bids arrive periodically, with values that are i.i.d. positive random variables. When a bid arrives, the seller should decide whether to accept it or not, weighting the possibility of the arrival of a better offer against the cost of waiting, which consists of periodic holding costs, discounting of future revenues and expenditures, and advertising. Despite the complex model, rather simple structural results are attainable. We present a simple dynamic programming calculation scheme of the optimal double-argument policy (when to sell and how to advertise).
Original language | English |
---|---|
Article number | 125836 |
Journal | Applied Mathematics and Computation |
Volume | 395 |
DOIs | |
State | Published - 15 Apr 2021 |
Externally published | Yes |
Keywords
- Decision analysis
- Dynamic programming
- Marketing
ASJC Scopus subject areas
- Computational Mathematics
- Applied Mathematics