Ant　colony　optimization　(ACO)　is　a　kind　of　simulated　evolutionary　algorithm.　It　imitates　ants＇　foraging　process　to　find　the　shortest　path,　coexists　with　the　characteristics　of　randomness　and　heuristic.　It　is　applied　successfully　to　solve　combinatorial　optimization　problems,　such　as　the　TSP　problem,　the　job-shop　scheduling　problem,　etc.　In　practical　application,　ACO　has　the　limitation　of　easily　being　trapped　into　local　optimum　and　long　time　to　converge.　We　propose　an　improved　ant　colony　optimization　algorithm,　consisting　of　introducing　random　factor　and　introducing　elitist　ants　as　well　as　weakened　strategy.　Random　factor　provides　a　direction　to　search　within　the　field　of　the　optimal　path.　Elitist　ants　and　weakened　strategy　strengthens　the　pheromone　above　the　shortest　path　and　weakens　the　pheromone　above　the　suboptimal　path　to　decrease　the　accumulated　impact.　Both　of　them　shorten　the　convergence　time.　Simulation　results　show　that　the　improved　algorithm　has　a　better　performance　than　the　traditional　one.　It　can　not　only　find　a　shorter　path　but　also　cost　less　convergence　time,　along　with　satisfactory　time　complexity.　The　best　path　length　of　TSPlib　pr136　we　get　is　96910,　closed　to　official　record　96772　and　relative　error　is　0:14%.　©　2016　IEEE.