A Flexible Branch and Bound Algorithm for Solving a Class of Linear Multiplicative Programming Problems
This paper presents a branch and bound algorithm with a flexible branching rules for a class of linear multiplicative programming(LMP)problems.Firstly,the(LMP)problem is transformed into an equivalent problem,and then the concave part of the nonconvex constraint is approximated by piecewise linear approximation.By using the proposed adaptive branching rule for dividing rectangles and iteratively refining the piecewise linear approximations,the solving process of the(LMP)problem is transformed into solving a series of second order cone relaxation problems(SOCR).In addition,the convergence and complexity of the algorithm are proved.Finally,numerical results show the effectiveness and feasibility of the proposed algorithm.
Linear multiplicative programmingGlobal optimal solutionFlexible branch and boundSecond order cone relaxation
万方数据
维普
