Morphological methods for design of modular systems (a survey)

The article addresses morphological approaches to design of modular systems. The following methods are briefly described: (i) basic version of morphological analysis (MA), (ii) modification of MA as method of closeness to ideal point(s), (iii reducing of MA to linear programming, (iv) multiple choice problem, (v) quadratic assignment problem, (vi) Pareto-based MA (i.e., revelation of Pareto-efficient solutions), (vii) Hierarchical Morphological Multicriteria Design (HMMD) approach, and (viii) Hierarchical Morphological Multicriteria Design (HMMD) approach based on fuzzy estimates. The above-mentioned methods are illustrated by schemes, models, and illustrative examples. An additional realistic example (design of GSM network) is presented to illustrate main considered methods.

💡 Research Summary

The paper presents a comprehensive survey of morphological methods that have been developed for the design of modular systems. It begins with a concise description of the classic Morphological Analysis (MA) technique, which decomposes a system into a set of components (or functions) and enumerates all possible combinations of design alternatives in a “possibility matrix.” Recognizing the combinatorial explosion inherent in the basic approach, the authors discuss several extensions that embed constraints (e.g., compatibility, budget) and objective functions (e.g., cost, performance, reliability) directly into the analysis.

The first extension is the “closeness to ideal point” method. Here a multi‑dimensional ideal point—representing the best attainable values for each objective—is defined, and each candidate configuration is evaluated by its distance to this point (typically Euclidean). The distance‑minimization problem can be reformulated as a 0‑1 linear programming model: binary variables indicate the selection of a particular alternative for each component, while constraints enforce “exactly‑one‑choice” per component and compatibility relations. This reformulation enables the use of powerful LP solvers for relatively large design spaces.

Next, the paper links morphological analysis to well‑known combinatorial optimization problems. The Multiple‑Choice Problem (MCP) captures the “one‑alternative‑per‑component” requirement and is solved by minimizing (or maximizing) a weighted sum of selected alternatives. The Quadratic Assignment Problem (QAP) adds interaction costs between pairs of components, making it especially suitable for network‑type designs where inter‑module interference or communication cost matters. Although QAP is NP‑hard, the authors note that heuristic and meta‑heuristic approaches (genetic algorithms, simulated annealing, etc.) can produce high‑quality solutions in practice.

A Pareto‑based version of MA is introduced for truly multi‑objective scenarios. By computing a vector of objective values for each configuration, the method identifies the Pareto‑efficient frontier—solutions that are not dominated on all criteria. This frontier serves as a reduced, decision‑maker‑friendly set from which further preference articulation can be performed.

The Hierarchical Morphological Multicriteria Design (HMMD) framework constitutes the core of the survey. HMMD treats a modular system as a hierarchy (tree) of subsystems. At each node, a local set of alternatives is evaluated using multicriteria decision‑making (MCDM) techniques such as Analytic Hierarchy Process (AHP), TOPSIS, or VIKOR. Compatibility matrices capture inter‑level constraints, and the overall system score is obtained by aggregating the local scores in a bottom‑up manner. The authors also present a fuzzy‑logic extension of HMMD, where expert judgments expressed in linguistic terms (“high,” “medium,” “low”) are converted into fuzzy numbers. Fuzzy aggregation then yields a possibility (compatibility) degree for each configuration, allowing the designer to handle uncertainty and vagueness inherent in early‑stage engineering judgments.

To illustrate the practical relevance of each method, the paper provides a series of schematic examples followed by a realistic case study: the design of a GSM cellular network. The case study addresses four interrelated design tasks—site selection, frequency allocation, transmission path planning, and equipment choice. For each task the authors apply all eight morphological approaches:

1. Basic MA enumerates all feasible site‑frequency‑equipment combos, highlighting the explosion of possibilities.
2. Ideal‑point closeness quickly identifies a configuration that simultaneously minimizes deployment cost and maximizes coverage, thanks to the linear‑programming reformulation.
3. MCP formulation enforces the rule that each cell must receive exactly one frequency band, yielding a cost‑effective allocation.
4. QAP model captures interference between neighboring base stations; the resulting solution minimizes total interference while preserving coverage.
5. Pareto‑based MA produces a set of 12 non‑dominated solutions across the three objectives (cost, coverage, reliability), giving planners a clear trade‑off landscape.
6. HMMD structures the network design hierarchically (region → site → equipment) and uses AHP to rank alternatives at each level, resulting in a coherent, top‑down configuration.
7. Fuzzy HMMD incorporates expert statements such as “high reliability, medium cost, high capacity” and translates them into fuzzy scores; the method narrows the candidate set to three configurations that best satisfy the fuzzy criteria.
8. Linear‑programming reduction (a sub‑case of the ideal‑point method) demonstrates computational efficiency for large‑scale instances.

The comparative analysis shows that each method has distinct strengths: distance‑to‑ideal and LP reduction excel in speed; QAP captures interaction effects; Pareto analysis clarifies trade‑offs; HMMD provides a systematic hierarchical workflow; fuzzy HMMD handles imprecise expert knowledge.

In the concluding discussion the authors argue that morphological methods, especially when integrated with classical optimization and multicriteria decision‑making tools, constitute a versatile toolbox for modular system design. They emphasize that hierarchical and fuzzy extensions are particularly valuable for large, complex engineering projects where uncertainty, multiple stakeholders, and conflicting objectives are the norm. The survey thus positions morphological analysis not as a relic of the 1960s but as an evolving methodology that can be adapted to contemporary engineering challenges such as telecommunications, aerospace, and modular product families.