Evolutionary　algorithms　and　swarm　intelligence　ones　are　commonly　used　to　solve　many　complex　optimization　problems　in　different　fields.　Yet,　some　of　them　have　limited　performance　when　dealing　with　high-dimensional　complex　problems　because　they　often　require　enormous　computational　resources　to　yield　desired　solutions,　and　some　of　them　may　easily　trap　into　local　optima.　To　solve　this　problem,　this　work　proposes　a　Self-adaptive　Teaching-learning-based　Optimizer　with　an　improved　Radial　basis　function　model　and　a　sparse　Autoencoder　(STORA).　In　STORA,　a　Self-adaptive　Teaching-Learning-Based　Optimizer　(STLBO)　is　designed　to　dynamically　adjust　parameters　for　balancing　exploration　and　exploitation　abilities.　Then,　a　sparse　autoencoder　(SAE)　is　adopted　as　a　dimension　reduction　method　to　compress　a　search　space　into　a　lower-dimensional　one　for　more　efficiently　guiding　a　population　to　converge　towards　global　optima.　Besides,　an　Improved　Radial　Basis　Function　model　(IRBF)　is　designed　as　a　surrogate　one　to　balance　training　time　and　prediction　accuracy.　It　is　adopted　to　save　computational　resources　for　improving　overall　performance.　In　addition,　a　dynamic　population　allocation　strategy　is　adopted　to　well　integrate　SAE　and　IRBF　in　STORA.　We　evaluate　STORA　by　comparing　it　with　several　state-of-the-art　algorithms　through　eight　benchmark　functions.　We　further　test　its　actual　performance　by　applying　it　to　solve　a　real-world　computation　offloading　problem.