Dynamic single-machine scheduling under distributed decision-making

This paper presents a deterministic dynamic model for the single-machine scheduling problem. The model uses forecasts of future job arrivals with the current data to extract job interactions over time, updating the information on a rolling basis. The model is implemented in a distributed structure with both the machine and the jobs involved in decision-making to create a schedule. The decision-making is modelled similar to an auction with a theoretical basis for problem decomposition, bid construction and bid evaluation. Numerical results indicate that the model outperforms other distributed implementations in both static and dynamic implementations for a wide range of single-machine scheduling problems.