Programmable controller and machine learning device

A programmable controller which executes a plurality of sequence programs for controlling at least one machine included in a machining system and which changes execution time allocation to the sequence programs for each operation cycle in each operation stage of the machining system, the programmable controller comprising:

• a time allocation setting section for setting the execution time allocation;
• a stage analysis section for analyzing the operation stage of the machining system based on data regarding an operation state of the machine;
• and
• a machine learning device which has learned changing of the execution time allocation to the sequence programs for each operation cycle, wherein
• the machine learning device includes
• a state observation section for observing, as a state variable representing a current environmental state, execution time allocation data representing execution time allocation set at time of execution of the operation stage of the machining system, operation stage data representing the operation stage, and machine operation pattern data, the machine operation pattern data being data regarding the operation state of the machine in the operation stage,
• a learning section which has learned the changing of the execution time allocation to the sequence programs for each operation cycle in relation to the operation stage of the machining system and an operation pattern of the machine, and
• a decision-making section for making a decision regarding the changing of the execution time allocation to the sequence programs for each operation cycle, based on the state variable observed by the state observation section and a learning result of the learning section.

A machine learning device which is provided in a programmable controller that executes a plurality of sequence programs for controlling at least one machine included in a machining system and which learns changing of execution time allocation to the sequence programs for each operation cycle in each operation stage of the machining system, the machine learning device comprising:

• a state observation section for observing, as a state variable representing a current environmental state, execution time allocation data representing execution time allocation set at time of execution of the operation stage of the machining system, operation stage data representing the operation stage, and machine operation pattern data, the machine operation pattern data being data regarding the operation state of the machine in the operation stage;
• a determination data acquisition section for acquiring cycle time determination data for determining whether cycle time taken to execute the operation stage executed based on the set execution time allocation is appropriate or inappropriate, as determination data representing an appropriate/inappropriate determination result for changing of the execution time allocation to the sequence programs for each operation cycle;
• and
• a learning section for learning the changing of the execution time allocation to the sequence programs for each operation cycle in relation to the operation stage of the machining system and an operation pattern of the machine using the state variable and the determination data.