Pricing Strategies of AI-enabled and Regular Products

Yinmeng Li1, Zhaojun Yang1,*, Jun Sun2, Xu Hu1

2022 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM) | 978-1-6654-8687-3/22/$31.00 ©2022 IEEE | DOI: 10.1109/IEEM55944.2022.9989923

1
School of Economics and Management, Xidian University, Xi’an, Shaanxi, China
2
College of Business & Entrepreneurship, University of Texas Rio Grande Valley, Edinburg, TX, USA
([email protected])

Abstract - At present, the trend of artificial intelligence research on how to optimize the decision-making.
(AI) for product innovation is becoming widespread, and Furthermore, AI-enabled products and regular products are
many manufacturers enhance their products with AI. In the substitutable to some extent. When manufacturers price
process of AI-enabled product production and sales, AI-enabled products, they need to consider the retail price
manufacturers must make decisions on not only the retail of regular products in the market as well. In particular,
prices of AI-enabled products but also whether to reprice
their original products. This study investigates the optimal
whether it is necessary to reprice regular products when AI-
decisions for manufacturers, considering the degree of enabled products are introduced is unclear.
consumers’ AI preference and the substitutability of the two This study investigates the optimal pricing and AI-
products. We find that manufacturers often need to reprice level strategies for AI-enabled products considering
their regular products after launching AI-enabled products consumers’ AI preference, price sensitivity, and cross-
to gain maximum benefits. In addition, this study also price sensitivity. The numerical analysis answers the
explores the impact of price sensitivity and cross-price following questions: (1) How do consumers’ AI preference
sensitivity on the manufacturer's optimal pricing decisions. and price sensitivities affect manufacturers’ pricing of AI-
The results of numerical analyses yield managerial enabled products? (2) When manufacturers launch AI-
implications for product manufacturers engaged in AI
enhancement.
enabled products, do they want to reprice regular products?
The findings inform the pricing decisions of manufacturers
Keywords - artificial intelligence (AI), AI-enabled that make AI-enabled products in addition to regular
products, product pricing, technology level, AI preference products.

I. INTRODUCTION II. RESEARCH BACKGROUND

Artificial intelligence (AI) is defined as a set of AI-enabled products are different from smart products
powerful technologies that enables an object to interact in the market, which emphasize the upgrade of regular
with the environment and solve problems autonomously products by AI, whereas smart products are considered a
[1]. Due to its rapid growth in recent years, AI has been a fundamental change in the concept of regular products [5].
significant driving force of product innovation. The " AI + However, there is no clear definition and division of such
hardware" model has led to the birth of new products (i.e., advanced technology-based products in the extant
AI-enabled products). Many manufacturers use AI to literature, all of which denote products with intelligent
enhance their regular products, and these upgraded features [6]. Terms like smart objects [7], intelligent
products bring new services and experiences for users. products [8], smart products [6], and AI-enabled products
Such AI-enabled products as smart-home appliances, [9] are often used interchangeably [10].
autonomous cars, and voice-controlled devices are Most studies on such products are empirical in nature.
becoming increasingly popular as they bring great One found smartness as a key differentiator in the smart
convenience to life and work. product market impactful on customer happiness [11].
Compared to regular products, AI-enabled products Another showed that the product quality (e.g., device
require manufacturers to invest in intelligent functions and quality, functional quality) of AI-enabled smartwatches
decide their sophistication levels. For example, the Society significantly improves the satisfaction of users [12]. It is
of Automotive Engineers (SAE) International established further indicated that the environmental benefits of AI-
the five levels of driving automation [2]. Automakers enhanced products have a positive impact on consumers'
integrated AI with their automobiles and the most purchase decisions [13]. The findings suggest that
intelligent vehicles available on the market today reach just consumers pay attention to the technology level of these
one level below the top [3]. Similarly, AI-enabled products products. Therefore, pricing AI-enabled products
of the same category on e-commerce platforms like undoubtedly needs to consider the impact of the attribute
Amazon target different consumer segments with different of technology level on market demand.
levels of intelligent functions [4]. Therefore, manufacturers Based on the pricing of general products, extended
must decide intelligence level and product pricing when pricing research considers certain attributes of innovative
they launch AI-enabled products, but there is a lack of products, such as greenness (i.e., eco-friendly degree) [14-

0859

Authorized licensed use limited to: NORTHWESTERN POLYTECHNICAL UNIVERSITY. Downloaded on December 29,2022 at 14:47:11 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of The 2022 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM)

17]. One study examines the pricing of green products enabled and regular products with 𝑝 , 𝑝 , 𝑇 . The
based on the demand function [16]. Such a demand manufacturer's profit can be formulated as:
function can be extended to the pricing of green and regular
products with some substitutability [14, 15]. The extant 𝜋 = (𝑝 − 𝑐 )𝐷 +(𝑝 − 𝑐 )𝐷 − 𝜑𝑇 . (3)
research provides helpful hints on the pricing strategy of
AI-enabled products, as regular and AI-enabled products TABLE Ⅰ
are substitutable based on common underlying LIST OF NOTATIONS
functionalities. Nevertheless, there is a lack of pricing
research in the context of AI-enabled product supply chains. Symbol Definition
To fill the literature gap, this study develops a pricing 𝐷 The market demand of the regular products
model for AI-enabled products, which factors in the 𝐷 The market demand of the AI-enabled products
hardware and software attributes of products in terms of 𝛼 The total market potential demand
technology level to enhance the demand function. 𝛾 The regular products’ share of the total market potential
demand
𝛽 Price sensitivity of consumers
𝛿 The degree of consumers’ AI preference
III. MODEL DEVELOPMENT 𝜇 The cross-price sensitivity coefficient
𝑝 The retail price of regular products
A. Assumption and Notations 𝑝 The retail price of AI-enabled products
𝑐 The unit cost of regular products
𝑐 The unit cost of basic functions of AI-enabled products
Before constructing the model, it is necessary to make (hardware)
several assumptions. (1) Only one manufacturer and 𝜑 The technology cost coefficient
consumers are considered in the model. The manufacturer 𝑇 The technology level of AI-enabled products (software)
invests in AI to enhance their regular products and decides 𝜋 The manufacturer's profit in case 𝑖 (𝑖=1, 2)
technology level and retail price. Consumers decide
whether to buy the AI-enabled product according to its
price. (2) The manufacturer's production is equal to the IV. RESULTS AND DISCUSSIONS
market demand. The market demand for both products is
greater than zero. Consumers have the same price First, the optimal price and technology level of AI-
sensitivity for two products. (3) These two products differ enabled products are obtained in case 1. Then, the optimal
only in price and technology attributes. (4) 𝜑𝑇 is the prices of both AI-enabled and regular products are
investment made by the manufacturer to implement the AI calculated in case 2. Finally, the manufacturer’s optimal
technology[18]. 𝑇 is the technology level of a product, profits in these two cases are compared. The Appendix
and 𝜑 denotes the technology cost coefficient. Because of gives mathematical proofs of lemmas and propositions.
the complexity of technology development process, there
is likely a quadratic relationship between development cost A. Case 1 Optimal Price for AI-enabled Products
and technology level. In addition, the hardware of AI-
enabled products would also require to be improved, so the The optimal price of AI-enabled products is explored
unit hardware cost of AI-enabled products is greater than in this case under the known retail price of regular products.
the unit cost of regular products (i.e., 𝑐 > 𝑐 ). (5) The Solving the first-order derivative of the manufacturer's
impact of a product’s price on its demand is always greater profit function yields the equilibrium results, which are
than the impact of the substitute products on demand. That summarized in Lemma 1. Besides, we analyze the impact
is, consumers are more sensitive to the price of a product of consumers’ AI preference on manufacturers’ optimal
than to that of its substitute (i.e., 𝛽 > 𝜇). decisions in Proposition 1.
Based on the above assumptions, the demand functions
can be derived. Table 1 lists the main symbols used in Lemma 1. Under the condition of 4𝛽𝜑 − 𝛿 > 0 , the
modeling. optimal price of AI-enabled products (𝑝 ∗ ), the optimal
technology level (𝑇 ∗ ) , the optimal profit of
𝐷 = 𝛾𝛼 − 𝛽𝑝 + 𝜇𝑝 (1) manufacturers (𝜋 ∗ ) are respectively:
𝐷 = (1 − 𝛾)𝛼 − 𝛽𝑝 + 𝛿𝑇 + 𝜇𝑝 (2)
∗
(2𝛽𝜑 − 𝛿 )𝑐 + 2𝜑[𝛼(1 − 𝛾) + 𝜇(2𝑝 − 𝑐 )]
𝑝 = (4)
B. Benchmark Model 4𝛽𝜑 − 𝛿
𝛿[𝛽𝑐 − 𝛼 + 𝛼𝛾 − 𝜇(2𝑝 − 𝑐 )]
The manufacturer sells regular and AI-enabled 𝑇 ∗=− (5)
4𝛽𝜑 − 𝛿
products, considering whether to reprice regular products 𝛿 (𝑐 − 𝑝 )(𝛼𝛾 − 𝛽𝑝 + 𝑐 𝜇) − 2𝜑𝛽𝜇𝑐 𝑐
or not. Therefore, two models are constructed. One case is +𝜑𝛽 (4𝑐 𝑝 + 𝑐 − 4𝑝 )+𝛼 𝜑(𝛾 − 1)
that manufacturers do not reprice regular products, and
+(𝛾 − 1)[2𝛼𝛽𝜑𝑐 + 2𝜑𝛼𝜇(𝑐 − 2𝑝 )]
make decisions only for AI-enabled products with the
decision variables 𝑝 , 𝑇. Another case is to price both AI- +𝜑𝜇 (𝑐 − 2𝑝 ) + 4𝛾𝛼𝛽𝜑(𝑝 − 𝑐 )
𝜋∗ = (6)
4𝛽𝜑 − 𝛿

0860

Authorized licensed use limited to: NORTHWESTERN POLYTECHNICAL UNIVERSITY. Downloaded on December 29,2022 at 14:47:11 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of The 2022 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM)

Proposition 1: The impact of the consumers’ AI preference Proposition 2 indicates that the price 𝑝 ∗ and the
on product pricing without repricing regular products: technology level 𝑇 ∗ of AI-enabled products are also
∗ ∗
> 0, > 0. increasing functions of AI preference degree under
repricing the regular products. It yields the same
managerial insights as Proposition 1. In addition, the
Proposition 1 indicates that the price 𝑝 ∗ and the proposition also brings an interesting understanding that
technology level 𝑇 ∗ of AI-enabled products are the price of regular products is positively correlated with
monotonically increasing functions of consumers’ AI consumers’ AI preference. Therefore, as consumers’ AI
preference without repricing the regular products. This is preference increases, the manufacturer that produces both
consistent with our expectations. A greater degree of types of products will increase the price of regular products.
consumers’ AI preference increases the technology level. In general, the sales of AI-enabled products will encroach
This is because when the degree of consumers’ AI on the market share of regular products, undermining their
preference becomes larger, its impact on demand will profit margins. The manufacturer will raise the price of
increase. Then the manufacturer is likely to invest more regular products to compensate for market share loss. In the
capital to enhance the technology level of AI-enabled market, low-income consumers still have to choose regular
products. Their price will increase correspondingly, as products that are cheaper than AI-enabled products.
manufacturers want to transfer the investment cost of AI Therefore, the manufacturer will profit from increasing
technology to consumers. As for consumers, when their AI regular product prices to a reasonable extent.
preference is higher, they are also willing to pay a higher
purchase price for AI-enabled products. Therefore, the C. Comparison of Two Cases
manufacturer can benefit from increasing the price of AI-
enabled products. The comparison of the manufacturer’s optimal profits
across the above two cases leads to Lemma 3.
B. Case 2 Optimal Prices for Two Products
Lemma 3: As ∆𝜋 ≥ 0, the manufacturers always reprice
The manufacturer considers repricing regular products the regular products.
after the launch of AI-enabled products. Solving the first-
order derivative of the manufacturer's profit function yields Only when the optimal price of the regular products in
the equilibrium results, which are summarized in Lemma case 2 is equal to the original price, the profit difference is
2. In addition, Proposition 2 analyzes the impact of equal to 0, and the two strategies yield the same profit for
consumers’ AI preference on manufacturers’ optimal the manufacturer. In most cases, however, it is the optimal
decisions. decision for the manufacturer to reprice the regular
products. Given manufacturers are benefit-oriented, they
Lemma 2. Under the condition of 𝛽𝛿 − 4𝜑𝛽 + 4𝜑𝜇 < tend to reprice regular products when AI-enabled products
0 , the optimal price of regular products (𝑝 ∗ ) , the are introduced.
optimal price of AI-enabled products (𝑝 ∗ ), the optimal
technology level (𝑇 ∗ ) , the optimal profit of
manufacturers (𝜋 ∗ ) are as follows respectively: V. NUMERICAL STUDY

4𝑐 𝜑(𝜇 − 𝛽 ) + (𝑐 𝜇 + 𝛼𝛾 + 𝛽𝑐 )𝛿 This section focuses on the case of repricing regular

+4𝛼𝜑(𝜇𝛾 − 𝜇 − 𝛽𝛾) products because the manufacturer always maximizes its
𝑝 ∗= (7)
2𝛽𝛿 + 8𝜑𝜇 − 8𝜑𝛽 profitability in this case. Through numerical analyses, the
2𝜑𝑐 (𝜇 − 𝛽 ) + 𝛽(𝑐 𝛿 − 2𝛼𝜑) impacts of parameters such as price sensitivity and cross-
+2𝛼𝛾𝜑(𝛽 − 𝜇) price sensitivity on optimal pricing decisions can be
𝑝 ∗= (8)
𝛽𝛿 + 4𝜑𝜇 − 4𝜑𝛽 examined. Based on related studies [15, 19], the parameters
𝛿(−𝛼𝛽 + 𝛽 𝑐 − 𝑐 𝜇 + 𝛼𝛽𝛾 − 𝛼𝜇𝛾) are set as follows: 𝛼 = 100, 𝜑 = 2, 𝑐 = 14, 𝑐 = 17,
𝑇 ∗= (9)
−4𝜑𝛽 + 𝛽𝛿 + 4𝜑𝜇 𝛾 = 0.4.
4𝜑(𝛽𝜇 − 𝛽 )(𝑐 + 𝑐 ) − 4𝜑𝛽𝛼 The first part of the numerical analysis shows the
+8𝜑𝜇[(𝛾 − 1)𝛼 + 𝜇𝑐 ](𝛼𝛾 + 𝜇𝑐 ) relationship between the manufacturer's optimal decisions
+[𝛼𝛾(𝑐 − 𝑐 ) + 𝜇𝑐 𝑐 ]8𝜑𝛽 + 8𝛼𝑐 𝜑𝛽 and price sensitivity, as shown in Fig. 1. Under the setting
−8𝜑𝛽(𝛾 − 𝛾)𝛼 + 𝛿 (𝛼𝛾 − 𝛽𝑐 + 𝜇𝑐 ) of basic parameters, we assume that 𝛿 = 0.5, 𝜇 = 0.25，
𝜋∗ = (10) 𝛽 ∈ [0.3, 1].
4(−4𝜑𝛽 + 𝛽𝛿 + 4𝜑𝜇 )

Proposition 2: The effect of the degree of consumers’ AI

preference on product pricing with repricing regular
∗ ∗ ∗
products: > 0, > 0, > 0.

0861

Authorized licensed use limited to: NORTHWESTERN POLYTECHNICAL UNIVERSITY. Downloaded on December 29,2022 at 14:47:11 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of The 2022 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM)

VI. CONCLUSION

The findings contribute to the literature in both theory

and practice. The modeling of two cases explores the
optimal pricing of AI-enabled products that maximizes the
manufacturer’s profitability. Numerical analyses examine
the impacts of price sensitivity and cross-price sensitivity
on the manufacturer's decision-making. The results yield
useful suggestions for manufacturers that launch AI-
enabled products:
(1) Regardless of whether regular products are repriced, a
higher degree of consumers’ AI preference justifies the
increase of investment in the technology level.
Accordingly, the price of AI-enabled products can be
raised. Therefore, the manufacturer should assess the
degree of AI preference of consumers before making
Fig. 1. The impact of price sensitivity on retail price
decisions.
(2) After the manufacturer launches AI-enabled products,
Fig. 1 shows that the manufacturer’s optimal price
it always needs to reprice regular products to maximize
decreases as price sensitivity increases, but gradually
profitability. As the degree of consumers’ AI preference
stabilizes. In general, consumers in the market can be
increases, the manufacturer may raise the price of regular
segmented according to their price sensitivity. Consumers
products. In addition, the price sensitivity of consumers to
in the low and mid-range are more sensitive to price. When
AI-enabled products and substitutable regular products
targeting these consumers, the manufacturer can not invest
also affects pricing strategies. Therefore, manufacturers
too much in AI and avoid overpricing AI-enabled products.
need to take all these factors into account when making
pricing decisions.
This study has limitations that point to future research
directions. First of all, this study simplifies the model by
considering only one manufacturer. The optimal pricing
decisions with the participation of multiple organizations
in the supply chain can be further investigated. In addition,
this study only considers the impact of technology level on
product demand. Future studies may address other factors
like product quality and the interaction between hardware
and software.

ACKNOWLEDGMENT

This work was supported by the National Social

Science Fund of China (No. 20BGL045)
Fig. 2. The impact of cross-price sensitivity on retail price

The second part of the numerical analysis shows the REFERENCES

relationship between the manufacturer's optimal price and
cross-price sensitivity, as shown in Fig. 2. Under the setting [1] Y. B. Wang, P. Zheng, T. Peng, H. Y. Yang, and J. Zou,
of basic parameters, we assume that 𝛿 = 0.5, 𝛽 = 0.5, "Smart additive manufacturing: Current artificial
𝜇 ∈ [0, 0.45]. intelligence-enabled methods and future perspectives,"
Science China-Technological Sciences, vol. 63, no. 9, pp.
Fig. 2 shows that Manufacturers' optimal decisions all 1600-1611, 2020.
increase as cross-price sensitivity increases and the price [2] H. A. RESEARCH, "What Is a Self-Driving Car?," 2020,
gap between the two types of products is gradually Available:
decreasing. This suggests that manufacturers that produce https://www.caranddriver.com/research/a31996016/what-is-
both regular and AI-enabled products need to consider the a-self-driving-car/.
impact of price sensitivity of consumers to substitute [3] U. S. E. P. Agency, "Self-driving Vehicles," 2022, Available:
https://www.epa.gov/greenvehicles/self-driving-vehicles.
products when setting the price gap between two types of
products.

0862

Authorized licensed use limited to: NORTHWESTERN POLYTECHNICAL UNIVERSITY. Downloaded on December 29,2022 at 14:47:11 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of The 2022 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM)

[4] S. Gunasekar, S. Ray, S. K. Dixit, and M. R. Pa, "AI-enables APPENDIX

product purchase on Amazon: what are the consumers
saying?," foresight, in press. Proof of lemma 1. The Hessian matrix for case 1 is given
[5] Z. Mani and I. Chouk, "Drivers of consumers' resistance to by (11). When satisfying 4𝛽𝜑 − 𝛿 > 0 , the Hessian
smart products," Journal of Marketing Management, vol. 33,
matrix is negative. We have the equilibrium results in
no. 1-2, pp. 76-97, 2017.
[6] S. Raff, D. Wentzel, and N. Obwegeser, "Smart Products: lemma 1 using the first order conditions, = 0, =
Conceptual Review, Synthesis, and Research Directions*," 0.
Journal of Product Innovation Management, vol. 37, no. 5,
pp. 379-404, 2020. −2𝛽 𝛿
[7] T. P. Novak and D. L. Hoffman, "Relationship journeys in 𝐻 = (11)
𝛿 −2𝜑
the internet of things: a new framework for understanding
interactions between consumers and smart objects," Journal
of the Academy of Marketing Science, vol. 47, no. 2, pp. 216- Substituting the optimal results in lemma 1 to (3), the
237, 2019. manufacturer's optimal profit can be derived without
[8] S. A. Rijsdijk, E. J. Hultink, and A. Diamantopoulos, repricing the regular product as (6).
"Product intelligence: its conceptualization, measurement
and impact on consumer satisfaction," Journal of the Proof of lemma 2 can be proved as proof of lemma 1.
Academy of Marketing Science, vol. 35, no. 3, pp. 340-356,
2007. Proof of proposition 1. Taking the first derivative of the
[9] S. L. Du and C. Y. Xie, "Paradoxes of artificial intelligence
optimal results with respect to 𝛿 in lemma 1, we obtain
in consumer markets: Ethical challenges and opportunities,"
Journal of Business Research, vol. 129, pp. 961-974, 2021. ∗
[10] M. Rokonuzzaman, K. Kim, K. K. Dugar, and J. Fox, "What 𝜕𝑝 4𝜑𝛿[𝛼(1 − 𝛾) − 𝛽𝑐 − 𝜇𝑐 + 2𝑝 𝜇]
= (12)
makes an object smart? Conceptualization, development, and 𝜕𝛿 (4𝜑𝛽 − 𝛿 )
∗
validation of a scale to measure the Smartness of a Thing 𝜕𝑇 [𝛼(1 − 𝛾) − 𝛽𝑐 + 𝜇(−𝑐 + 2𝑝 )](4𝜑𝛽 + 𝛿 )
(SoT)," Journal of Business Research, vol. 141, pp. 337-354, = (13)
𝜕𝛿 (4𝜑𝛽 − 𝛿 )
2022.
[11] B. Henkens, K. Verleye, and B. Lariviere, "The smarter, the
Recall that 𝑇 ∗
> 0, 4𝛽𝜑 − 𝛿 > 0 , we have 𝛼(1 −
better?! Customer well-being, engagement, and perceptions ∗
in smart service systems," International Journal of Research 𝛾) − 𝛽𝑐 − 𝜇𝑐 + 2𝜇𝑝 > 0 . Then we have > 0,
in Marketing, vol. 38, no. 2, pp. 425-447, 2021. ∗
[12] M. U. H. Uzir et al., "Applied Artificial Intelligence and user > 0.
satisfaction: Smartwatch usage for healthcare in Bangladesh
during COVID-19," Technology in Society, vol. 67, p. Proposition 2 can be proved as proof of proposition 1.
101780, 2021.
[13] B. Frank, "Artificial intelligence-enabled environmental
sustainability of products: Marketing benefits and their Proof of lemma 3. The difference between the two cases
variation by consumer, location, and product types," Journal is
of Cleaner Production, vol. 285, p. 125242, 2021. ∆𝜋 = 𝜋 ∗ − 𝜋 ∗
[14] Z. Basiri and J. Heydari, "A mathematical model for green −𝐴
supply chain coordination with substitutable products," ∆𝜋 = (14)
4(4𝛽𝜑 − 𝛿 )(𝛽𝛿 + 4𝜑𝜇 − 4𝜑𝛽 )
Journal of Cleaner Production, vol. 145, pp. 232-249, 2017.
[15] P. Ma, C. Zhang, X. P. Hong, and H. Xu, "Pricing decisions
for substitutable products with green manufacturing in a Where 𝐴 = 4𝜑(𝜇 − 𝛽 )(𝑐 − 2𝑝 ) + 𝛿 (𝑐 𝜇 + 𝛼𝛾 +
competitive supply chain," Journal of Cleaner Production, 𝛽𝑐 − 2𝛽𝑝 ) + 4𝛼𝜑𝜇(𝛾 − 1) − 4𝛼𝜑𝛽𝛾.
vol. 183, pp. 618-640, 2018. Recall that (4𝛽𝜑 − 𝛿 ) > 0, (−4𝜑𝛽 + 𝛽𝛿 +
[16] D. Ghosh and J. Shah, "Supply chain analysis under green 4𝜑𝜇 ) < 0 . We have (4𝛽𝜑 − 𝛿 )(−4𝜑𝛽 + 𝛽𝛿 +
sensitive consumer demand and cost sharing contract," 4𝜑𝜇 ) < 0. Therefore, we obtain ∆𝜋 ≥ 0.
International Journal of Production Economics, vol. 164, pp.
319-329, 2015.
[17] S. Saha, I. Nielsen, and S. S. Sana, "Effect of Optimal
Subsidy Rate and Strategic Behaviour of Supply Chain
Members under Competition on Green Product Retailing,"
Mathematical Problems in Engineering, vol. 2021, p.
6661103, 2021.
[18] L. Zhao, R. Y. Zhong, and Y. Zhang, "Cooperative models
for technology supply chain in online social community,"
Computers & Industrial Engineering, vol. Vol.145, p.
106460, 2020.
[19] J. Zhao, J. Wei, and Y. J. Li, "Pricing decisions for
substitutable products in a two-echelon supply chain with
firms' different channel powers," International Journal of
Production Economics, vol. 153, pp. 243-252, 2014.

0863

Authorized licensed use limited to: NORTHWESTERN POLYTECHNICAL UNIVERSITY. Downloaded on December 29,2022 at 14:47:11 UTC from IEEE Xplore. Restrictions apply.