Provided is a job plan creation system capable of creating an efficient job plan when the job plan is reconfigured on the basis of the progress of the jobs after the completion of the jobs that day and by the job starting time of the following day, for example. A job plan creation system extracts an interrupted job that has been started but has not been completed, a delayed job that should have been started based on the job plan but has not been started, and a not-started job that is scheduled to be started on the job plan and so has not been started from the job plan, and creates the job plan while setting priorities in the order of the interrupted job, the delayed job and the not-started job.

The disclosed is a system and production processes, comprising a processor to perform the steps: a) obtaining a long-term production schedule containing production operations for a set of production orders by using forecasted values for production operation durations; b) deriving dispatching recommendations for suitable order release dates, dispatching sequences and/or resource choices of all subsequent operational dispatching decisions from the obtained long-term production schedule; c) obtaining a released subset of the set of production orders having each a release date and a due date and including information of all possible dispatching sequences and logical dependencies between them to produce a requested product; d) iteratively calculating subsequent operational dispatching decisions based on operational constraints and on the derived dispatching recommendations which support a decision-making in favor of optimizing the overall production process with respect to long-term objectives; e) outputting each calculated operational dispatching decision to a production executer.

An electronic device and a method of filtering dispatching rules for a workstation are provided. The method includes the following. A dispatch rule set including a first dispatching rule and a first test data set are obtained. An initial individual is generated according to the dispatch rule set based on a genetic algorithm. A first individual is generated according to the first test data set and the initial individual based on the genetic algorithm, and the first individual includes a first dispatching rule sequence corresponding to a first workstation, and the first dispatching rule sequence includes a first dispatching rule. A first score corresponding to the first dispatching rule is calculated according to a first order of the first dispatching rule in the first dispatching rule sequence, and the dispatching rule set is updated according to the first score. The dispatching rule set that is updated is output.

Provided is a dispatching method in a factory based on reinforcement learning. The dispatching method in a factory based on reinforcement learning may comprise: constructing a Markov decision process (MDP) for dispatching actions of a dispatcher in the factory and resulting rewards and states of the factory; performing learning by applying reinforcement learning (RL) to the constructed MDP; and as a result of said RL, selecting a job that maximizes a weighted sum of a plurality of scored dispatching rules.

A system includes a dispatching system and at least one processor. The dispatching system is configured to provide dispatching preferences indicating that tools feedback preferences thereof for at least one lot. The processor is coupled to the dispatching system and at least one memory. Computer program code encoded in the memory is executed by the processor to cause the system to process the lot using at least one of the tools with utilization of a tool-lot relationship that is generated based on the dispatching preferences. A non-transitory computer readable medium and a method are also disclosed herein.

Provided is a dispatching method in a factory based on reinforcement learning. The dispatching method in a factory based on reinforcement learning may comprise: constructing a Markov decision process (MDP) for dispatching actions of a dispatcher in the factory and resulting rewards and states of the factory; performing learning by applying reinforcement learning (RL) to the constructed MDP; and as a result of said RL, selecting a job that maximizes a weighted sum of a plurality of scored dispatching rules.

A method is disclosed and includes generating a tool-lot relationship based on at least one process constraint and a dispatching rule which includes dispatching preferences for operating at least one lot; assigning the at least one lot to at least one tool by using the tool-lot relationship; and the at least one tool, assigned using the tool-lot relationship, processing the at least one lot by utilization of the tool-lot relationship to manufacture semiconductor devices on at least one wafer in the at least one lot.

A system includes a dispatching system and at least one processor. The dispatching system is configured to provide dispatching preferences indicating that tools feedback preferences thereof for at least one lot. The processor is coupled to the dispatching system and at least one memory. Computer program code encoded in the memory is executed by the processor to cause the system to process the lot using at least one of the tools with utilization of a tool-lot relationship that is generated based on the dispatching preferences. A non-transitory computer readable medium and a method are also disclosed herein.

A method is disclosed and includes generating a tool-lot relationship based on at least one process constraint and a dispatching rule which includes dispatching preferences for operating at least one lot; assigning the at least one lot to at least one tool by using the tool-lot relationship; and the at least one tool, assigned using the tool-lot relationship, processing the at least one lot by utilization of the tool-lot relationship to manufacture semiconductor devices on at least one wafer in the at least one lot.

A method is disclosed and includes determining whether there is any WIP information with one or more process constraints in a process constraint database that is coupled with a manufacturing execution system; determining whether there is a dispatching rule in the dispatching rule database that is coupled with a dispatching system; generating a tool-lot relationship based on at least one of the process constraints and the dispatching rule; utilizing the tool-lot relationship to assign one or more lots to one or more tools respectively.