模拟退火算法来源于固体退火原理，将固体加温至充分高，再让其徐徐冷却，加温时，固体内部粒子随温升变为无序状，内能增大，而徐徐冷却时粒子渐趋有序，在每个温度都达到平衡态，最后在常温时达到基态，内能减为最小。根据Metropolis准则，粒子在温度T时趋于平衡的概率为e-ΔE/(kT)，其中E为温度T时的内能，ΔE为其改变量，k为Boltzmann常数。用固体退火模拟组合优化问题，将内能E模拟为目标函数值f，温度T演化成控制参数t，即得到解组合优化问题的模拟退火算法：由初始解i和控制参数初值t开始，对当前解重复“产生新解→计算目标函数差→接受或舍弃”的迭代，并逐步衰减t值，算法终止时的当前解即为所得近似最优解，这是基于蒙特卡罗迭代求解法的一种启发式随机搜索过程。退火过程由冷却进度表(Cooling

Schedule)控制，包括控制参数的初值t及其衰减因子Δt、每个t值时的迭代次数L和停止条件S。 -Simulated annealing algorithm derived from the theory of solid annealing, the solid heat to full high and let it slowly cooling, heating, the temperature rise inside the solid particles with the shape into disorder, which can be increased gradually while slowly cooling particles increasingly ordered, the temperature has reached equilibrium in each state, and finally reached the ground state at room temperature, which can be reduced to minimum. According to Metropolis criterion, particles tend to equilibrium at a temperature T, the probability e-ΔE/(kT), where E is the temperature T, internal energy, ΔE change its volume, k the Boltzmann constant. Simulated annealing with a solid portfolio optimization problem, the internal energy E is modeled as the objective function value f, temperature T evolved into control parameter t, which are solutions of combinatorial optimization problems of the simulated annealing algorithm: the initial solution from the initial value of t i and the control