Back to search

Democritus University of Thrace (DUTH)

Intelligent methodologies for the assessment and selection of industrial robotic systems and technologies

Abstract

dc:description

The expanding diversity of industrial robots, each with unique technical capabilities, cost structures, and operational profiles, presents a significant challenge for modern manufacturing. This dissertation addresses this challenge by proposing, analyzing, and comparing methodologies that refine robot evaluation and selection to better align with industrial goals and constraints. It begins with a comprehensive, integrated literature review that categorizes existing robot selection methods, including Multi-Criteria Decision Making (MCDM) models, Production System Performance Optimization (PSPO) models, computer-aided systems, and soft computing tools. Empirical analysis through systematic benchmarking revealed that traditional static ranking models often produce inconsistent and divergent rankings, demonstrating their insufficiency in capturing objective reality and highlighting the critical need for more robust approaches. Building on these insights, the dissertation introduces a novel Fuzzy Digraph model, designed to overcome the limitations of static models and the inherent uncertainties in expert judgments. This framework seamlessly integrates qualitative expert judgments, such as user-friendliness or maintainability, with quantitative performance data, like payload or cost, by representing all inputs as triangular fuzzy numbers and processing them with fuzzy arithmetic. A significant innovation is its ability to handle attribute interdependencies by systematically computing the interrelation matrix directly from the normalized average weights assigned to each attribute, thereby drastically simplifying the decision-making process and reducing the cognitive load on experts compared to complex pairwise comparisons. The model translates expert opinions into fuzzy linguistic variables, constructs directed graphs to encode relationships among attributes, and aggregates this information into a single, streamlined “selection index,” providing a holistic view of each candidate robot’s suitability. Furthermore, recognizing that relationships between robot attributes are often nonlinear, the research makes a fundamental theoretical contribution by advancing the modeling capabilities of Fuzzy Cognitive Maps (FCMs). It moves beyond conventional linear influence models by developing and evaluating a flexible, generalized nonlinear influence function based on the logistic function. This allows for a more realistic representation of cause-effect relationships, capturing complex dynamics such as saturation effects, threshold effects, and varying growth rates of influence. Empirical evaluations demonstrate that this nonlinear approach enhances the model’s dynamic behavior and achieves stable equilibrium states even when conventional linear models fail or fall into non-converging limit cycles. This advancement is a domain-agnostic contribution to computational intelligence and complex systems modeling.The culmination of this work is the synthesis of the Fuzzy Digraph and nonlinear FCM models into two advanced, hybrid cognitive decision-making frameworks. The first, the New Fuzzy Digraph (NFD), enhances the original model by employing an FCM as a feedback mechanism to refine initial expert judgments on attribute weights, adjusting them based on simulated interdependencies. The second, the FCM-based Decision-Making (FCMDM) model, represents a pure FCM approach that treats the entire decision problem as a dynamic system; robot attributes are set as initial conditions, and the optimal choice is derived from the emergent, systemic outcome of the simulation of a dedicated output node. These frameworks transform the role of the human expert from a source of static inputs to an integral part of a collaborative, human-in-the-loop cognitive system, enabling powerful "what-if" scenario analysis. Numerical case studies confirm that both methods yield stable and adaptable rankings, accommodating a broad range of operational demands and reducing the risk of overlooking critical interdependencies among robot attributes. Collectively, these contributions mark a paradigm shift from static multi-criteria ranking to dynamic, complex systems modeling in industrial robot selection. By integrating fuzzy digraph techniques, advanced nonlinear FCMs, and classic MCDM principles into cohesive selection procedures, this dissertation provides a robust foundation for organizations to select robotic solutions that truly reflect their performance targets, economic realities, and evolving operational needs in an ever-advancing industrial landscape.

Degree

thesis:*
Grantor dc:publisher
Democritus University of Thrace (DUTH)
Year dc:date
2025

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Ketipi, Maria
  • Κετίπη, Μαρία

Subjects

dc:subject × 16

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/60910

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Ketipi, Maria; Κετίπη, Μαρία. Intelligent methodologies for the assessment and selection of industrial robotic systems and technologies. Democritus University of Thrace (DUTH), 2025. http://hdl.handle.net/10442/hedi/60910