Search-Based Software Testing: Using Affordable Computing Power for Efficient and Effective Software Testing

Computing power is increasingly cheap and readily-available, whether on our laptops, in the cloud, or even in novel architectures such GPU cards. The premise of Search-Based Software Testing (SBST) is that this computing power can and should be used to assist in the software testing process: the objective is to use the computing power to automate the most tedious, labour-intensive parts of testing so that the human engineers can apply their experience to higher-level tasks where it is most valuable. For example, SBST techniques are effective in identifying and removing redundant test cases, optimising the order of tests so that defects are found earlier, and generating highly-efficient test inputs. In this talk, I will explain the general approach of SBST, demonstrate some SBST techniques from my own research, and argue why -- as software systems become increasingly large and complex -- the degree of automation that SBST enables is vital for efficient and effective testing.