Bellow you can find our papers regarding applications in Marketing
Designing a successful product line is essential for companies aiming to satisfy diverse consumer preferences while remaining competitive in fast-paced markets. By systematically determining the right combination of product attributes, brands can optimize objectives such as profit, market share, or sustainability. The following works highlight how advanced optimization techniques can address the formidable complexity of the product line design problem.
The paper of Zervoudakis et al. (2020) merges the strengths of Firefly Algorithm and Genetic Algorithm to achieve better accuracy and faster convergence. By adopting the swarm-intelligence principles of Firefly and pairing them with the well-known genetic operations of crossover and mutation, this hybrid approach manages to avoid local optima. Comparative evaluations reveal improved exploration of large solution spaces and enhanced product-line configurations that better reflect consumer preference data.
The paper of Tsafarakis et al. (2022) explores how Tabu Search, a long-established local-search methodology, can outperform several well-known heuristics in complex design scenarios. By maintaining a dynamic memory of forbidden or unproductive moves, Tabu Search efficiently navigates vast product configuration spaces. The paper also proposes a specialized variant that further reduces runtime, allowing for faster turnarounds in real-world settings where product attribute possibilities may number in the millions.
The paper of Tsafarakis et al. (2020) shows how Differential Evolution, a powerful stochastic optimizer, can be tailored to discrete, combinatorial tasks using fuzzy logic. This self-tuning approach adjusts crucial operators at every iteration, significantly improving solution diversity and convergence speed. Thorough comparisons with classical differential evolution variants and other metaheuristics highlight the consistency and flexibility of this fuzzy self-tuning framework.
The paper of Pantourakis et al. (2022) conducts an extensive investigation of biologically inspired approaches for shaping product lines. Drawing on immune system principles, Clonal Selection Algorithms iteratively refine candidate solutions through clone generation and hypermutation. Across multiple data sets of different scales, the paper compares numerous algorithmic variants, clarifying how specialized operators drive performance. The findings demonstrate that CSA can deftly balance local search intensification with global exploration, outperforming or matching many traditional heuristics.
The paper of Zervoudakis & Tsafarakis (2024) adapts the Bees Algorithm to the combinatorial complexities of product design and integrates fuzzy logic to automatically calibrate algorithmic parameters. This eliminates the need for trial-and-error parameter settings and further refines the search for high-quality solutions. Experiments confirm that this fusion of fuzzy self-tuning with swarm-based intelligence can reliably pinpoint effective product combinations, matching or exceeding prior heuristic methods.
The paper of Tsafarakis & Zervoudakis (2024) examines how multi-objective optimization can balance profit with social and ecological benefits. The methodology leverages a Non-dominated Sorting Genetic Algorithm to incorporate both financial and welfare objectives, demonstrating how product lines can be efficiently reconfigured to incorporate suboptimal products that ordinarily would be discarded. This helps firms minimize unnecessary waste while still meeting the market’s quality perceptions.
References
Pantourakis, M., Tsafarakis, S., Zervoudakis, K., Altsitsiadis, E., Andronikidis, A., & Ntamadaki, V. (2022). Clonal Selection Algorithms for Optimal Product Line Design: A Comparative Study. European Journal of Operational Research, 298(2), 585–595. https://doi.org/10.1016/J.EJOR.2021.07.006
Tsafarakis, S., & Zervoudakis, K. (2024). Optimal Product Line Re-Design for Reducing Food Wasted due to Marketing Standards. In G. K. D. Saharidis (Ed.), Proceedings of the International Conference on “Energy, Sustainability and Climate Crisis” 2024 (pp. 113–117). http://escc.uth.gr/wp-content/uploads/2025/01/ESCC-2024_Book-of-Proceedings.pdf
Tsafarakis, S., Zervoudakis, K., & Andronikidis, A. (2022). Optimal product line design using Tabu Search. Journal of the Operational Research Society, 73(9), 2104–2115. https://doi.org/10.1080/01605682.2021.1954486
Tsafarakis, S., Zervoudakis, K., Andronikidis, A., & Altsitsiadis, E. (2020). Fuzzy self-tuning differential evolution for optimal product line design. European Journal of Operational Research, 287(3), 1161–1169. https://doi.org/10.1016/j.ejor.2020.05.018
Zervoudakis, K., & Tsafarakis, S. (2024). Fuzzy Self-tuning Bees Algorithm for designing optimal product lines. Applied Soft Computing, 167, 112228. https://doi.org/10.1016/J.ASOC.2024.112228
Zervoudakis, K., Tsafarakis, S., & Paraskevi-Panagiota, S. (2020). A New Hybrid Firefly – Genetic Algorithm for the Optimal Product Line Design Problem. In N. Matsatsinis, Y. Marinakis, & P. Pardalos (Eds.), Learning and Intelligent Optimization. LION 2019. Lecture Notes in Computer Science (pp. 284–297). Springer, Cham. https://doi.org/10.1007/978-3-030-38629-0_23
Customer segmentation is a fundamental practice in modern marketing, allowing businesses to categorize consumers based on shared traits such as demographics, behavioral patterns, or spending habits. By forming these groups, firms can craft more relevant product offers and target their advertising strategies more effectively, ultimately boosting customer satisfaction and loyalty.
In our paper of Zervoudakis & Tsafarakis (2025) we employ a novel bio-inspired optimizer that draws on the survival patterns of flying foxes under extreme heat. The proposed method autonomously adjusts its internal parameters—thanks to fuzzy self-tuning—while continuously refining cluster centroids in a vast, multi-dimensional data space. Empirical evaluations, including both simulated and real-world datasets, reveal that this flying fox-based approach yields more robust and interpretable customer segments than classical clustering methods such as k-means. Moreover, because it automatically tunes its parameters, the algorithm overcomes traditional clustering pitfalls like reliance on initial random solutions or a predetermined number of segments. Consequently, businesses can discover clearer, more actionable groupings of customers, leading to data-driven marketing campaigns that resonate better with each segment’s unique profile and behavior.
References
Zervoudakis, K., & Tsafarakis, S. (2025). Customer segmentation using flying fox optimization algorithm. Journal of Combinatorial Optimization, 49(1), 1–20. https://doi.org/10.1007/S10878-024-01243-6