Machine learning devices and methods for optimizing the speed and accuracy of thread mill, inner diameter, outer shape, and surface machine tools

A machine learning method of a machine learning device configured to perform machine learning with respect to a numerical control device that operates a machine tool on the basis of a machining program,

• the machine learning method performed by at least one processor of the machine learning device, the machine learning method comprising:
• acquiring state information including a setting value, a cycle time of cutting a workpiece, and a machining accuracy of the workpiece, the setting value including a spindle speed, a feed rate, a number of cuts, a cutting amount per one time, and a tool compensation amount;
• selecting action information including modification information of a setting value of at least the number of cuts, the cutting amount per one time, and the tool compensation amount included in the state information based on a measure of selecting an action with which the cutting amount per one time is increased and/or the number of cuts are decreased, and outputting the action information thus selected to the numerical control device;
• calculating a reward value in reinforcement learning based on the cycle time and the machining accuracy included in the state information;
• and
• updating an action value function on the basis of the reward value that has been calculated, the state information, and the action information, wherein
• the spindle speed, the feed rate, the number of cuts, the cutting amount per one time, and the tool compensation amount are set in a predetermined machining program,
• the numerical control device causes the machine tool to perform cutting by executing the machining program, and
• the cycle time and the machining accuracy are acquired by the numerical control device executing the machining program.