Historical Information Stability based Reward for Reinforcement Learning in Continuous Integration Testing.
Tiange CaoZheng LiRuilian ZhaoYang YangPublished in: QRS (2021)
Keyphrases
- reinforcement learning
- historical information
- integration testing
- black box
- object oriented programs
- action space
- state space
- software testing
- eligibility traces
- function approximation
- reinforcement learning algorithms
- learning algorithm
- decision making
- markov decision processes
- component based software
- multi agent
- test cases
- temporal difference
- dynamic programming
- optimal policy
- policy gradient
- model free
- partially observable environments
- control policy
- long run
- reward function
- source code
- data sets
- learning agent
- average reward
- state action
- database
- address these issues
- software engineering
- machine learning