# 创建优化训练任务
[如何为智能体匹配最佳奖惩机制](https://support.huaweicloud.com/ops-agentarts/agentarts_14_0105.html)
:
[步骤一：配置优化对象与模型](https://support.huaweicloud.com/ops-agentarts/agentarts_14_0106.html)
:
[步骤二：配置优化方法和奖惩机制](https://support.huaweicloud.com/ops-agentarts/agentarts_14_0107.html)
:
[步骤三：配置超参数](https://support.huaweicloud.com/ops-agentarts/agentarts_14_0108.html)
:
[步骤四：绑定数据集、产物路径与委托](https://support.huaweicloud.com/ops-agentarts/agentarts_14_0109.html)
:
[附录：自适应奖励payload数据格式说明](https://support.huaweicloud.com/ops-agentarts/agentarts_14_0141.html)
:
