Abstract
Product Portfolio Planning (PPP) is one of the most critical decisions for companies to gain an edge in the competitive market. It seeks for the optimal combination of products and attribute levels offered for customers in the target market, which is a NP-hard combinatorial optimization problem. In this paper, a Probability-based Discrete Particle Swarm Optimization (PDPSO) algorithm is proposed to solve the PPP problem. In PDPSO, the particle is encoded as discrete values, which can be straightforwardly used to represent the product portfolio with discrete attributes. PDPSO adopts a probability-based mechanism to update particles. Specifically, a probability vector is used to decide the probability of three search behaviors, i.e., learning from the personal best position, global best position, or random search. In experiments, the search performance of PDPSO has been compared with that of a Genetic Algorithm (GA) and a Simulated Annealing (SA) algorithm on generated PPP problem cases with different sizes. The results indicate that PDPSO obtains significantly better optimization results than GA and SA in most cases and obtains desirable/near-optimal solutions on various PPP problem cases. A case study of notebook computer portfolio planning is also presented to illustrate the efficiency and effectiveness of PDPSO.
Similar content being viewed by others
Data availability
The data of this paper has been uploaded to Mendeley, a cloud database belonging to Elsevier. It can be downloaded from the following website. https://doi.org/10.17632/tbm8tv27z7.1
References
Albritton MD, McMullen PR (2007) Optimal product design using a colony of virtual ants. Eur J Oper Res 176(1):498–520
Alexouda G (2004) An evolutionary algorithm approach to the share of choices problem in the product line design. Comput Oper Res 31(13):2215–2229
Alexouda G (2005) A user-friendly marketing decision support system for the product line design using evolutionary algorithms. Decis Support Syst 38:495–509
Alexouda G, Paparrizos K (2001) A genetic algorithm approach to the product line design problem using the seller’s return criterion: an extensive comparative computational study. Eur J Oper Res 134(1):165–178
Ali SI, Ali A, AlKilabi M, Christie M (2021) Optimal supply chain design with product family: a cloud-based framework with real-time data consideration. Comput Oper Res 126:105112
Balakrishnan PV, Jacob VS (1995) Triangulation in decision support systems: algorithms for product design. Decis Support Syst 14:313–327
Balakrishnan PVS, Jacob VS (1996) Genetic algorithms for product design. Manag Sci 42(8):1105–1117
Balakrishnan PV, Gupta R, Jacob VS (2004) Development of hybrid genetic algorithms for product line designs. Syst Man Cybern 34(1):468–483
Balakrishnan PV, Gupta R, Jacob VS (2006) An investigation of mating and population maintenance strategies in hybrid genetic heuristics for product line designs. Comput Oper Res 33(3):639–659
Belloni A, Freund R, Selove M, Simester D (2008) Optimizing product line designs: efficient methods and comparisons. Manag Sci 54(9):1544–1552
D’Souza B, Simpson TW (2003) A genetic algorithm based method for product family design optimization. Eng Optim 35(1):1–18
Das S, Mullick SS, Suganthan PN (2016) Recent advances in differential evolution–an updated survey. Swarm Evol Comput 27:1–30
Dash B, Gajanand MS, Narendran TT (2018) A model for planning the product portfolio and launch timings under resource constraints. Int J Prod Res 56(15):5081–5103
Du G, Xia Y, Jiao RJ, Liu X (2019) Leader-follower joint optimization problems in product family design. J Intell Manuf 30(3):1387–1405
ElMaraghy H, Schuh G, ElMaraghy W, Piller F, Schönsleben P, Tseng M, Bernard A (2013) Product variety management. CIRP Ann 62(2):629–652
Gauss L, Lacerda DP, Cauchick Miguel PA (2021) Module-based product family design: systematic literature review and meta-synthesis. J Intell Manuf 32(1):265–312
Gharehchopogh FS, Gholizadeh H (2019) A comprehensive survey: whale optimization algorithm and its applications. Swarm Evol Comput 48:1–24
Goswami M, Pratap S, Kumar SK (2016) An integrated Bayesian-Game theoretic approach for product portfolio planning of a multi-attributed product in a duopolistic market. Int J Prod Res 54(23):6997–7013
Green PE, Krieger AM (1985) Models and heuristics for product line selection. Mark Sci 4(1):1–19
Hsieh FS (2020) A comparative study of several metaheuristic algorithms to optimize monetary incentive in ridesharing systems. ISPRS Int J Geo Inf 9(10):590
Hsieh FS (2022) Development and comparison of ten differential-evolution and particle swarm-optimization based algorithms for discount-guaranteed ridesharing systems. Appl Sci 12(19):9544
Huffman C, Kahn BE (1998) Variety for sale: Mass customization or mass confusion? J Retail 74(4):491–513
Jiao J, Zhang Y (2005) Product portfolio planning with customer-engineering interaction. IIE Trans 37(9):801–814
Jiao J, Zhang Y, Wang Y (2007) A heuristic genetic algorithm for product portfolio planning. Comput Oper Res 34(6):1777–1799
Jiao JR, Simpson TW, Siddique Z (2007) Product family design and platform-based product development: a state-of-the-art review. J Intell Manuf 18(1):5–29
Kaul A, Rao VR (1995) Research for product positioning and design decisions: an integrative review. Int J Res Mark 12(4):293–320
Kennedy J, Eberhart R (1995) Particle swarm optimization. In Proceedings of ICNN'95-international conference on neural networks (Vol. 4, pp. 1942–1948). IEEE
Kohli R, Krishnamurti R (1989) Optimal product design using conjoint analysis: computational complexity and algorithms. Eur J Oper Res 40(2):186–195
Kumar D, Chen W, Simpson TW (2009) A market-driven approach to product family design. Int J Prod Res 47(1):71–104
Li A-D, He Z (2020) Multiobjective feature selection for key quality characteristic identification in production processes using a nondominated-sorting-based whale optimization algorithm. Comput Ind Eng 149:106852
Li A-D, Xue B, Zhang M (2021) Improved binary particle swarm optimization for feature selection with new initialization and search space reduction strategies. Appl Soft Comput 106:107302
Li J, Wei X, Li B, Zeng Z (2022) A survey on firefly algorithms. Neurocomputing 500:662–678
Li AD, Liu X (2023) Synthetic datasets on the product portfolio planning (PPP) problem. Mendeley. https://doi.org/10.17632/TBM8TV27Z7.1
Liu Z, Wong YS, Lee KS (2011) A manufacturing-oriented approach for multi-platforming product family design with modified genetic algorithm. J Intell Manuf 22(6):891–907
Liu C, Yang X, Wang J (2021) Optimization of product line considering compatibility and reliability via discrete imperialist competitive algorithm. RAIRO-Oper Res 55(6):3773–3795
Luo L (2011) Product line design for consumer durables: an integrated marketing and engineering approach. J Mark Res 48(1):128–139
Matzler K, Waiguny M, Fuller J (2007) Spoiled for choice: consumer confusion in Internet-based mass customization. Innov Mark 3(3):7–18
Miao C, Du G, Jiao RJ, Zhang T (2017) Coordinated optimisation of platform-driven product line planning by bilevel programming. Int J Prod Res 55(13):3808–3831
Michalek JJ, Ceryan O, Papalambros PY, Koren Y (2006) Balancing marketing and manufacturing objectives in product line design. J Mech Des 128(6):1196–1204
Michalek JJ, Ebbes P, Adigüzel F, Feinberg FM, Papalambros PY (2011) Enhancing marketing with engineering: optimal product line design for heterogeneous markets. Int J Res Mark 28(1):1–12
Müller S, Haase K (2016) On the product portfolio planning problem with customer-engineering interaction. Oper Res Lett 44(3):390–393
Ogawa S, Piller FT (2006) Reducing the risks of new product development. MIT Sloan Manag Rev 47(2):65
Pantourakis M, Tsafarakis S, Zervoudakis K, Altsitsiadis E, Andronikidis A, Ntamadaki V (2021) Clonal selection algorithms for optimal product line design: a comparative study. Eur J Oper Res 298:585
Patterson ML (2005) New product portfolio planning and management. In: The PDMA handbook of new product development. 2:46–58
Raman N, Chhajed D (1995) Simultaneous determination of product attributes and prices, and production processes in product-line design. J Oper Manag 12:187–204
Sadeghi A, Alem-Tabriz A, Zandieh M (2011) Product portfolio planning: a metaheuristic-based simulated annealing algorithm. Int J Prod Res 49(8):2327–2350
Simpson TW, Seepersad CC, Mistree F (2001) Balancing commonality and performance within the concurrent design of multiple products in a product family. Concurr Eng 9(3):177–190
Steiner W, Hruschka H (2003) Genetic algorithms for product design: how well do they really work? Int J Mark Res 45(2):229–240
Stone RB, Kurtadikar RM, Villanueva N, Arnold CB (2008) A customer needs motivated conceptual design methodology for product portfolio planning. J Eng Des 19(6):489–514
Thevenot HJ, Simpson TW (2006) Commonality indices for product family design: a detailed comparison. J Eng Des 17(2):99–119
Thevenot HJ, Alizon F, Simpson TW, Shooter SB (2007) An index-based method to manage the tradeoff between diversity and commonality during product family design. Concurr Eng 15(2):127–139
Tsafarakis S (2016) Redesigning product lines in a period of economic crisis: a hybrid simulated annealing algorithm with crossover. Ann Oper Res 247(2):617–633
Tsafarakis S, Marinakis Y, Matsatsinis N (2011) Particle swarm optimization for optimal product line design. Int J Res Mark 28(1):13–22
Tsafarakis S, Saridakis C, Baltas G, Matsatsinis N (2013) Hybrid particle swarm optimization with mutation for optimizing industrial product lines: an application to a mixed solution space considering both discrete and continuous design variables. Ind Mark Manag 42(4):496–506
Tsafarakis S, Zervoudakis K, Andronikidis A, Altsitsiadis E (2020) Fuzzy self-tuning differential evolution for optimal product line design. Eur J Oper Res 287(3):1161–1169
Tsafarakis S, Zervoudakis K, Andronikidis A (2021) Optimal product line design using Tabu Search. J Oper Res Soc 73:2104
Xue B, Nguyen S, Zhang M (2014) A new binary particle swarm optimisation algorithm for feature selection. In: European Conference on the applications of evolutionary computation (pp. 501–513). Springer, Berlin, Heidelberg
Zervoudakis K, Tsafarakis S (2022) A global optimizer inspired from the survival strategies of flying foxes. Eng Comput. https://doi.org/10.1007/s00366-021-01554-w
Funding
This work was supported by the National Natural Science Foundation of China (NSFC) [Grant Number 72101182, 72231005]; and the Humanities and Social Sciences Youth Fund of Ministry of Education of China [Grant Numbers 19YJC630071, 19YJC630221].
Author information
Authors and Affiliations
Contributions
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by XL and ADL. The first draft of the manuscript was written by ADL and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors have no relevant financial or non-financial interests to disclose.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix A: Tuning the parameters for PDPSO
Appendix A: Tuning the parameters for PDPSO
In this section, we verify different combinations of parameters for the proposed PDPSO on the PPP problem with different sizes and obtain a best compromise setting for the PDPSO parameters. The obtained PDPSO parameters have been used in the case study in Sect. 5 and further search performance analysis in Sect. 6. The tuning experiments are run on a PC with a 3.4 gigahertz CPU and a 16 gigabytes RAM with Matlab R2021a.
The parameters used in PDPSO that may affect its search performance are swarm size \(N\), maximum number of iterations \(T\), and parameter \(p^{r}\) that controls the search behavior. In the tuning experiments, we set two candidate settings for \(N\) and \(T\). The first setting is \(N = 100\) and \(T = 200\). The second setting is \(N = 50\) and \(T = 400\). Both settings yield the same \(20,000\) function evaluations. The parameter \(p^{r}\) that controls the random search probability during the iterations is suggested to be a small value. In the tuning experiments, five values of \(p^{r}\) (i.e., 0.001, 0.005, 0.01, 0.05, and 0.1) are examined. Therefore, a total number of \(2 \times 5\) = 10 parameter combinations are yielded for \(N\), \(T\), and \(p^{r}\).
A strategy like that in Sect. 6.2 is used to generate the PPP problem cases for parameter tuning. Problem parameters to generate the PPP problem cases are set to be \(J \in \left\{ {5,8} \right\}\), \(\left( {K + 1} \right) \in \left\{ {6,9,12} \right\}\), \(L_{k} \in \left\{ {6,9,12} \right\}\), and \(I \in \left\{ {3,4,5} \right\}\), which are the same as that used in Sect. 6.2. This results in \(2 \times 3 \times 3 \times 3 = 54\) problem parameter combinations (please refer to Table 7 for the detailed description of each problem parameter combination). For each problem parameter combination, one problem case (which is smaller than that used in Sect. 6.2 considering limited computational resources) is generated and a total number of 54 problem cases (ID:1-54) are generated. Finally, the 54 problem cases and the 10 parameter combinations yield a total number of 540 tuning experiments.
For each of the 540 tuning experiments, the PDPSO algorithm is repeated 30 times to comprehensively evaluate the performance of PDPSO. Therefore, each of the tuning experiments would have 30 sets of the experimental results. The expected shared surplus \({\text{E}}\left[ V \right]\) as shown in Eq. (1) is the objective to be maximized by PDPSO and thus it is a good measure to evaluate the performance of PDPSO. Moreover, whether the PDPSO algorithm can obtain results with small standard deviation during the 30 runs is also a key measure to evaluate PDPSO since the standard deviation reflects the robustness of the search algorithm. Therefore, considering both the mean and standard deviation of the expected shared surplus \({\text{E}}\left[ V \right]\) over 30 runs, the signal-to-noise (S/N) ratio is used as the performance measure in order to find a desirable parameter setting for PDPSO as suggested in Sadeghi et al. (2011). The \({\text{E}}[V\)] is larger-the-better and thus the corresponding loss function is defined as
where \(N_{r}\) is the number of experimental runs (30 in these tuning experiments) and \(E\left[ V \right]_{r}\) is the expected shared surplus obtained from the \(r\) th run.
The S/N ratios obtained by the 10 PDPSO parameter settings on the generated 54 PPP problem cases are shown in Table
11. The parameter setting that obtains the best S/N ratio on each problem case is highlighted in bold, and the #best row of the table denotes the number of cases that a parameter setting obtains the best (highest) S/N ratio compared with other parameter settings. According to the table, it can be found that the parameter setting \(\left\{ {N = 50,\,{ }T = 400,{ }\,p^{r} = 0.01} \right\}\) obtains the highest S/N ratios in 33 of the 54 cases, which means that this parameter setting performs the best in most of the cases. According to the results, problem size seems not to be a key factor that affects the performance of the parameter setting \(\left\{ {N = 50,{ }\,\,T = 400,{ }\,p^{r} = 0.01} \right\}\) since this setting can obtain the best S/N ratio results on different problem sizes. It can be also found that the parameter \(p^{r}\) should be carefully set since a too high or a too low \(p^{r}\) value can significantly affect the results. If only the five parameter settings where \(N = 100\), \(T = 200\) and \(p^{r} \in \left\{ {0.001,\,\,{ }0.005,{ }\,\,0.01,\,\,{ }0.05,{ }\,\,{ }0.1} \right\}\) are considered, it can be found that the parameter setting \(N = 100\), \(T = 200\), and \(p^{r} = 0.01\) obtains the best S/N ratios in 32 of the 54 cases. This denotes that no matter the {\(N = 50, T = 400\}\) or \(\left\{ {N = 100, T = 200} \right\}\) is used, \(p^{r} = 0.01\) is a desirable setting for PDPSO. Overall, according to the S/N ratio results in Table 11, a desirable parameter setting (i.e., \(N = 50\), \({ }T = 400\), and \({ }p^{r} = 0.01\)) is obtained for PDPSO. This parameter setting is used in the case study and the search performance analysis sections.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Liu, X., Li, AD. An improved probability-based discrete particle swarm optimization algorithm for solving the product portfolio planning problem. Soft Comput 28, 2535–2562 (2024). https://doi.org/10.1007/s00500-023-08530-0
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00500-023-08530-0