Towards Four-Layer Framework of Combinatorial Problems
Mark Sh. Levin · 2008
Four-layer framework of combinatorial problems/models is suggested for applied problems structuring and solving: (1) basic combinatorial models and multicriteria DM-models (e.g., ranking, clustering, knapsack problem, multiple choice problem, multicriteria ranking, combinatorial synthesis, assignment/allocation); (2) composite models/procedures (e.g., multicriteria combinatorial problems); (3) basic (standard) solving frameworks, e.g.: (i) hierarchical morphological multicriteria design (HMMD) (ranking, combinatorial synthesis), (ii) multistage design (two-level HMMD), (iii) special multistage composite framework (clustering, assignment/location, multiple choice problem); and (4) domain-oriented solving frameworks, e.g.: (a) design of modular software, (b) design of test inputs for multi-function system testing, (c) combinatorial planning of medical treatment, (d) improvement of communication network. The multi-layer scheme covers 'decision cycle', i.e., problem statement, models, algorithms/procedures, solving schemes, decisions, decision analysis and improvement. The framework of problems corresponds to open complex systems.