The improved quantum-behaved particle swarm optimization (QPSO) with global criterion is presented to improve the convergence at the global-best point of current multi-dimensional (≥5) QPSO, and to judge whether the optimization result is the global-best point or not. Based on QPSO with self-adapting adjustment of the inertia weight, the positions of particles are periodically mutated, and the search performance of the swarm is mutated according to the evolution speed and aggregation degree. Meanwhile, the global criterion is established to judge optimization results by the aggregation degree factor. The algorithms and criterion are tested with benchmark functions and cone-derived waverider configuration optimization problems. The experiments show that the search performance of improved QPSO is significantly improved, and the results are true and reliable. The global criterion is practical and effective.