Very interesting study. I have some specific questions about your model.
The first question is about the winner determination problem. As shown on page 14, why is it necessary to maximise the sum of accepted bid values? If using a Dutch auction, intuitively I think that the allocation is automatically determined as the clock time proceeds.
The second question is about the agent behaviour rule. How do you model the bidder’s behaviour? Is it honest bidding? In the case of a Dutch auction, if an agent does not know other agents’ valuations, the agent is necessary to estimate other’s bidding values. Probably, honest bidding is not rational.
—
The proposed method allows customers to submit multiple bids at the same time.
If a customer makes multiple bids at a time, the restaurant optimizes a winner determination problem to determine which bids made by the customer will be accepted.
The proposed method solves the winner determination problem each $AT$ day.
The proposed method is strictly different from the Dutch auction.
—
A customer agent decides to book a restaurant on day $d$ and submit bids honestly on day $d$ immediately.
Customers do not wait to submit bids, even if there are still a lot of empty tables on day $d$.
(We conducted a questionnaire survey to define the distribution of a day $d$ that customers decide to book a restaurant.)
However, a human may wait to submit bids in that situation.
Conducting experiments with human subjects and analyzing human behavior are future directions.
Dear Mr. Suginouchi, thank you for your presentation of a study which is still quite appropriate due to the continuing persistence of the Covid-19 emergency. The possibility to reduce the waiting time for receiving confirmation of the custumer request is of very high relevance both specifically for the restaurant reservation management and for more general manufacturing contexts. Could you suggest some production engineering sector, such as supply chain management, where your approach could be effectively applied?
The proposed method aims to gather and integrate distributed knowledge and preferences among customers.
We focused on the restaurant reservation considering social distancing because restaurant reservation is suitable for Dutch auction, besides the table allocation problem is complex.
The proposed method can be adapted to not only restaurant reservations but also various situations where there are multiple customers.
– BtoC manufacturing company’s production scheduling (Bidder: Customers ordering a product)
– Transportation scheduling for AGV system (Bidder: Job or AGV)
– Inventory management for multi-level supply chain（Bidder: companies constituting SC）
The proposed method is valid for coordinating customers’ opinions and establishing a schedule that all subjects can accept.
Very interesting study. I have some specific questions about your model.
The first question is about the winner determination problem. As shown on page 14, why is it necessary to maximise the sum of accepted bid values? If using a Dutch auction, intuitively I think that the allocation is automatically determined as the clock time proceeds.
The second question is about the agent behaviour rule. How do you model the bidder’s behaviour? Is it honest bidding? In the case of a Dutch auction, if an agent does not know other agents’ valuations, the agent is necessary to estimate other’s bidding values. Probably, honest bidding is not rational.
Thank you very much for your questions.
—
The proposed method allows customers to submit multiple bids at the same time.
If a customer makes multiple bids at a time, the restaurant optimizes a winner determination problem to determine which bids made by the customer will be accepted.
The proposed method solves the winner determination problem each $AT$ day.
The proposed method is strictly different from the Dutch auction.
—
A customer agent decides to book a restaurant on day $d$ and submit bids honestly on day $d$ immediately.
Customers do not wait to submit bids, even if there are still a lot of empty tables on day $d$.
(We conducted a questionnaire survey to define the distribution of a day $d$ that customers decide to book a restaurant.)
However, a human may wait to submit bids in that situation.
Conducting experiments with human subjects and analyzing human behavior are future directions.
Dear Mr. Suginouchi, thank you for your presentation of a study which is still quite appropriate due to the continuing persistence of the Covid-19 emergency. The possibility to reduce the waiting time for receiving confirmation of the custumer request is of very high relevance both specifically for the restaurant reservation management and for more general manufacturing contexts. Could you suggest some production engineering sector, such as supply chain management, where your approach could be effectively applied?
Thank you very much for your question.
The proposed method aims to gather and integrate distributed knowledge and preferences among customers.
We focused on the restaurant reservation considering social distancing because restaurant reservation is suitable for Dutch auction, besides the table allocation problem is complex.
The proposed method can be adapted to not only restaurant reservations but also various situations where there are multiple customers.
– BtoC manufacturing company’s production scheduling (Bidder: Customers ordering a product)
– Transportation scheduling for AGV system (Bidder: Job or AGV)
– Inventory management for multi-level supply chain（Bidder: companies constituting SC）
The proposed method is valid for coordinating customers’ opinions and establishing a schedule that all subjects can accept.