Abstract: The optimization of beer recipe is a powerful approach to improve the efficiency of beer company.But for recipe optimization problems，the traditional mathematical optimization methods achieve more complex and lack an overall search for the optimal solution robustness.Ant Colony Algorithm（ACA） is fit to solve the combinatorial optimization problems，but it has disadvantage of slow convergence and time-consuming in the process of evolution.Therefore，the variable scale ant colony algorithm is presented and the scope of the search is shorted in the iterative process to improve the efficiency of optimization in the paper.Then the study of new ACA of the formulation of beer recipe is also presented，which in meeting production targets，and achieve the lowest total cost of the raw materials.Simulation results show that compared with the traditional ACA，the improved ACA has more global search capability and better robustness，also has practical value because of its easy implementation.

Key words: optimization, variable scale ant colony algorithm, global search, beer recipe

摘要： 啤酒配方优化是提高啤酒企业生产效率的重要途径。但对于配方优化问题，传统的数学优化方法实现较为复杂，缺乏全局最优解搜索的鲁棒性。蚁群算法目前多用于组合优化问题，但它在演化过程中有收敛慢、耗时长的缺点。因此，提出了变尺度蚁群算法，在迭代过程中不断收缩蚂蚁的搜索范围以提高优化效率。并研究了变尺度蚁群算法在啤酒配方优化中的应用，在满足生产指标前提下，实现配方的原料总成本最低。其应用结果表明：针对啤酒配方优化这类连续域问题，变尺度蚁群算法具有更强的全局搜索能力和鲁棒性，并易于实现，具有实际应用价值。

关键词: 优化, 变尺度蚁群算法, 全局搜索, 啤酒配方