The invention discloses a dynamic distributing method of
software testing resources based on a multi-object difference
evolutionary algorithm, wherein the method can be applied to modularized tests on serial /
parallel software systems. Aiming at the situations that the resources are wasted due to the fact that the existing test methods do not consider the variations along with the test progress, the method comprises the following steps: S1, encoding feasible solutions; S2, dividing test stages; S3, setting
control parameters; S4, initializing clusters; S5, implementing
mutation operations; S6, implementing interlace operations; S7, implementing selection operations based on non-dominate
ranking; S8, judging whether the maximum iteration time is reached, and
jumping to step S5 if it is not; otherwise,
jumping to step S9; S9, decoding pareto solutions; S10, judging whether non-implemented test stages still exist, and ending the solving process if not, or otherwise,
jumping to step S11; updating testing resources for the next test stage, and jumping to step S4 for entering the next test stage.