Abstract: This application discloses a method, an apparatus, and a device for scheduling virtual objects in a virtual environment, which relate to artificial intelligence (AI), and belong to the field of computer technologies. The method includes: obtaining frame data generated by an application program associated with the virtual environment; performing feature extraction on the frame data to obtain a state feature of a target virtual object; performing deduction on the state feature of the target virtual object according to N scheduling policies; invoking a value network prediction model to process the N subsequent state features, to obtain expected returns of executing N scheduling policies; and controlling the target virtual object to execute a scheduling policy having a highest expected return.

Abstract: This application discloses a method, an apparatus, and a device for scheduling virtual objects in a virtual environment, which relate to artificial intelligence (AI), and belong to the field of computer technologies. The method includes: obtaining frame data generated by an application program associated with the virtual environment; performing feature extraction on the frame data to obtain a state feature of a target virtual object; performing deduction on the state feature of the target virtual object according to N scheduling policies; invoking a value network prediction model to process the N subsequent state features, to obtain expected returns of executing N scheduling policies; and controlling the target virtual object to execute a scheduling policy having a highest expected return.

Abstract: This application discloses a method, an apparatus, and a device for scheduling virtual objects in a virtual environment, which relate to artificial intelligence (AI), and belong to the field of computer technologies. The method includes: obtaining frame data generated by an application program associated with the virtual environment; performing feature extraction on the frame data to obtain a state feature of a target virtual object; performing deduction on the state feature of the target virtual object according to N scheduling policies; invoking a value network prediction model to process the N subsequent state features, to obtain expected returns of executing N scheduling policies; and controlling the target virtual object to execute a scheduling policy having a highest expected return.

Abstract: This application discloses a method, an apparatus, and a device for scheduling virtual objects in a virtual environment, which relate to artificial intelligence (AI), and belong to the field of computer technologies. The method includes: obtaining frame data generated by an application program associated with the virtual environment; performing feature extraction on the frame data to obtain a state feature of a target virtual object; performing deduction on the state feature of the target virtual object according to N scheduling policies; invoking a value network prediction model to process the N subsequent state features, to obtain expected returns of executing N scheduling policies; and controlling the target virtual object to execute a scheduling policy having a highest expected return.