Optimal solution search device, device for controlling controlled object by optimizing algorithm, and optimal solution search program

An algorithm-based optimizing controller which optimizes control characteristics of a control system by searching for an optimal solution of an evaluation function using an optimizing algorithm while repeatedly evaluating output from the evaluation function which receives input of control coefficients that affect control characteristics of the control system for controlling characteristics of a controlled object and outputs the control characteristics, characterized in that:

• searches based on the optimizing algorithm are carried out using a combination of interactive evaluation method evaluating the output from the evaluation function based on interaction with a user and autonomous evaluation method evaluating the output from the evaluation function based on predetermined evaluation criteria;
• the controller comprises a first control system for controlling a first characteristic of the controlled object, and a second control system for controlling a second characteristic of the controlled object;
• the first control system searches for an optimal solution of the first evaluation function using the optimizing algorithm while repeatedly evaluating, by the autonomous evaluation method, the output from the first evaluation function which receives input of control coefficients that affect control characteristics of the first control system and outputs the control characteristics;
• and
• the second control system searches for an optimal solution of a second evaluation function using the optimizing algorithm while repeatedly evaluating, by the interactive evaluation method, the output from the second evaluation function such that the output from the first evaluation function will be within a predetermined range based on the output from the first evaluation function determined from the optimal solution found in the optimization of the first control system using the second evaluation function which receives input of control coefficients that affect control characteristics of the second control system and outputs the control characteristics as well as using the first evaluation function.

An optimal solution search program which is a computer-executable program for optimizing control characteristics of a first control system for controlling a first characteristic of a controlled object and a second control system for controlling a second characteristic of the controlled object, by searching for an optimal solution of an evaluation function using an optimizing algorithm while repeatedly evaluating output from the evaluation function which receives input of control coefficients that affect control characteristics of the control system for controlling characteristics of the controlled object and outputs the control characteristics, characterized in that the program makes the computer:

• carry out searches based on the optimizing algorithm using a combination of interactive evaluation method evaluating the output from the evaluation function based on interaction with a user and autonomous evaluation method evaluating the output from the evaluation function based on predetermined evaluation criteria;
• search for an optimal solution of a first evaluation function, in the first control system, using the optimizing algorithm while repeatedly evaluating, by the interactive evaluation method, the output from the first evaluation function which receives input of control coefficients that affect control characteristics of the first control system and outputs the control characteristics;
• and
• search for an optimal solution of a second evaluation function, in the second control system, using the optimizing algorithm while repeatedly evaluating, by the autonomous evaluation method, the output from the second evaluation function such that the output from the first evaluation function will be within a predetermined range based on the output from the first evaluation function determined from the optimal solution found in the optimization of the first control system using the second evaluation function which receives input of control coefficients that affect control characteristics of the second control system and outputs the control characteristics as well as using the first evaluation function.

An optimal solution search program which is a computer-executable program for optimizing control characteristics of a first control system for controlling a first characteristic of a controlled object and a second control system for controlling a second characteristic of the controlled object, by searching for an optimal solution of an evaluation function using an optimizing algorithm while repeatedly evaluating output from the evaluation function which receives input of control coefficients that affect control characteristics of the control system for controlling characteristics of the controlled object and outputs the control characteristics, characterized in that the program makes the computer:

• carry out searches based on the optimizing algorithm using a combination of interactive evaluation method evaluating the output from the evaluation function based on interaction with a user and autonomous evaluation method evaluating the output from the evaluation function based on predetermined evaluation criteria;
• search for an optimal solution of a first evaluation function, in the first control system, using the optimizing algorithm while repeatedly evaluating, by the autonomous evaluation method, the output from the first evaluation function which receives input of control coefficients that affect control characteristics of the first control system and outputs the control characteristics;
• and
• search for an optimal solution of a second evaluation function, in the second control system, using the optimizing algorithm while repeatedly evaluating, by the interactive evaluation method, the output from the second evaluation function such that the output from the first evaluation function will be within a predetermined range based on the output from the first evaluation function determined from the optimal solution found in the optimization of the first control system using the second evaluation function which receives input of control coefficients that affect control characteristics of the second control system and outputs the control characteristics as well as using the first evaluation function.