公开(公告)号：US10460360B2

公开(公告)日：2019-10-29

申请号：US14299925

申请日：2014-06-09

Applicant: SUNRUN, INC.

Inventor： Gary Wayne ,  Billy Hinners ,  John Hovell ,  Charles Buhler ,  Jake Wachman

IPC: G06F17/50 ,  G06Q30/06 ,  G06F9/445 ,  G06Q30/02

Abstract: A configuration engine traverses sequential levels of a decision tree in order to iteratively refine a configuration for a solar power system. At each level of the decision tree, the configuration engine determines the outcome of a design decision based on computing the result of a value function. The configuration engine explores configurations that optimize the value function result compared to other configurations, and may also discard less optimal configurations. When a current configuration is considered less optimal than a previous configuration generated at a previous level, the configuration engine discards the current configuration and re-traverses the decision tree starting with the previous configuration.

公开(公告)号：US20160004796A1

公开(公告)日：2016-01-07

申请号：US14854384

申请日：2015-09-15

Applicant: Sunrun, Inc.

Inventor： Billy Hinners ,  Gary Wayne

IPC: G06F17/50 ,  G06N5/04

CPC classification number: G06F17/5004 ,  F24S2201/00 ,  G06F17/50 ,  G06F17/5009 ,  G06F2217/08 ,  G06N5/045 ,  Y02E60/76 ,  Y04S40/22

Abstract: A configuration engine traverses sequential levels of a decision tree in order to iteratively refine a configuration for a solar power system. At each level of the decision tree, the configuration engine determines the outcome of a design decision based on computing the result of a value function. The configuration engine explores configurations that optimize the value function result compared to other configurations, and may also discard less optimal configurations. When a current configuration is considered less optimal than a previous configuration generated at a previous level, the configuration engine discards the current configuration and re-traverses the decision tree starting with the previous configuration.

Abstract translation: 配置引擎遍历决策树的顺序级别，以迭代地优化太阳能发电系统的配置。 在决策树的每个级别，配置引擎基于计算值函数的结果来确定设计决策的结果。 配置引擎会探索与其他配置相比优化值函数结果的配置，也可以丢弃较不优化的配置。 当当前配置被认为不如先前级别生成的先前配置的最优配置时，配置引擎将丢弃当前配置，并从先前配置开始重新执行决策树。