一种求解无单调性拟凸均衡问题的外梯度算法
An extragradient algorithm for quasiconvex equilibrium problems without monotonicity
AI总结:
针对拟凸均衡问题,提出一种以法次梯度为搜索方向、自适应更新步长的外梯度算法,在无单调性和 Lipschitz 条件下证明其收敛到 ρ-拟解或精确解。
摘要:
我们尝试提供一种算法,用于逼近 K. Fan 于1972年证明存在解的拟凸均衡问题的解。所提算法是一种迭代过程:每次迭代的搜索方向为法次梯度,同时步长更新无需满足 Lipschitz 型条件。当双函数 f 关于其第二变量半严格拟凸时,该算法对任意正 ρ 均收敛到一个 ρ-拟解;而当 f 强拟凸时,它收敛到问题的解。该方法既不要求单调性,也不要求 Lipschitz 性质。
英文摘要:
We attempt to provide an algorithm for approximating a solution of the quasiconvex equilibrium problem that was proved to exist by K. Fan 1972. The proposed algorithm is an iterative procedure, where the search direction at each iteration is a normal-subgradient, while the step-size is updated avoiding Lipschitz-type conditions. The algorithm is convergent to a $ρ$- quasi-solution with any positive $ρ$ if the bifunction $f$ is semistrictly quasiconvex in its second variable, while it converges to the solution when $f$ is strongly quasiconvex. Neither monotoniciy nor Lipschitz property is required.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。