Algebraic solution of tropical optimization problems via matrix sparsification with application to scheduling

作者：Krivulin Nikolai^*

来源：Journal of Logical and Algebraic Methods in Programming, 2017, 89: 150-170.

DOI：10.1016/j.jlamp.2017.03.004

• Tropical algebra
• Idempotent semifield
• Optimization problem
• Sparse matrix
• Backtracking
• Just-in-time scheduling

Optimization problems are considered in the framework of tropical algebra to minimize and maximize a nonlinear objective function defined on vectors over an idempotent semifield, and calculated using multiplicative conjugate transposition. To find the minimum of the function, we first obtain a partial solution, which explicitly represents a subset of solution vectors. We characterize all solutions by a system of simultaneous equation and inequality, and show that the solution set is closed under vector addition and scalar multiplication. A matrix sparsification technique is proposed to extend the partial solution, and then to obtain a complete solution described as a family of subsets. We offer a backtracking procedure that generates all members of the family, and derive an explicit representation for the complete solution. As another result, we deduce a complete solution of the maximization problem, given in a compact vector form by the use of sparsified matrices. The results obtained are illustrated with illuminating examples and graphical representations. We apply the results to solve real-world problems drawn from project (machine) scheduling, and give numerical examples.

• 出版日期2017-6