Reducing Separable Convex Programs with Tree Constraints
针对一类具有树约束的可分离凸规划问题，提出一种精简求解方法，通过求解最多2N个单变量凸子问题得到最优解，适用于大规模问题。
This paper describes a class of separable convex programs with tree constraints that has applications in production planning, quality improvement, and other related areas. A reduction procedure is presented for solving this class of separable convex programs with N variables. This reduction procedure determines an optimal solution to the convex problem by solving at most 2N simple convex subproblems with one variable. Hence, this reduction procedure is an efficient approach for solving large scale convex programs of this sort.