Abstract: Many practical problems can transformed to constrained optimization problems with various modality in real life and are all complex.It is difficult for a variety of traditionary methods based grads information to solve the constrained optimization problem with non-differential coefficient，multi-modal，nonconve and non-linear functions.But the current artificial algorithm shows excellence for above problems.After learning the outcome of differential evolution abroad，the paper uses improved differential evolution to solve constrained optimization problem.Lastly，the test examples are simulated to check up and the result shows superiority for differential evolution to solve constrained optimization problem.

Key words: differential evolution, constrained optimization, adaptation, testing

摘要： 在现实生活中许多实际问题都可以转化为约束优化问题，并且实际问题通常都很复杂，其函数形态各具特色，传统基于梯度信息的各种求解策略对于具有不可微、多峰及非凸的非线性函数约束优化问题很难凑效。而最近兴起的智能类算法却对这类问题的求解效果突出，在借鉴国外的差异演化算法研究成果基础上，运用改进差异演化算法来求解约束优化问题。最后通过实例进行仿真实验，结果表明改进差异演化算法在求解约束优化问题时具有一定的优越性。

关键词: 差异演化, 约束优化, 自适应, 测试