Mathematical Problems in Engineering

Mathematical Problems in Engineering / 2014 / Article
Special Issue

Stochastic Systems 2014

View this Special Issue

Research Article | Open Access

Volume 2014 |Article ID 539768 | 14 pages | https://doi.org/10.1155/2014/539768

Equilibrium Model of Discrete Dynamic Supply Chain Network with Random Demand and Advertisement Strategy

Academic Editor: Ming Gao
Received25 Apr 2014
Accepted17 Jul 2014
Published26 Aug 2014

Abstract

The advertisement can increase the consumers demand; therefore it is one of the most important marketing strategies in the operations management of enterprises. This paper aims to analyze the impact of advertising investment on a discrete dynamic supply chain network which consists of suppliers, manufactures, retailers, and demand markets associated at different tiers under random demand. The impact of advertising investment will last several planning periods besides the current period due to delay effect. Based on noncooperative game theory, variational inequality, and Lagrange dual theory, the optimal economic behaviors of the suppliers, the manufactures, the retailers, and the consumers in the demand markets are modeled. In turn, the supply chain network equilibrium model is proposed and computed by modified project contraction algorithm with fixed step. The effectiveness of the model is illustrated by numerical examples, and managerial insights are obtained through the analysis of advertising investment in multiple periods and advertising delay effect among different periods.

1. Introduction

In the 1980s, the interest in supply chain and supply chain management increased tremendously. Supply chain management, which incorporates the raw materials supplying, production and distribution in the demand markets in the end [1], is a hot topic in the academic world as well as the business community. There is abundance of research available on the supply chain management. We refer the readers to the work of [2] to achieve a comprehensive review on the supply chain topic.

These researches mainly focus on the stringy supply chain or a single manufacturer. In fact, the supply chain is a network which consists of suppliers, manufacturers, retailers, and demand markets [3]. Thus, there is limited contribution in the previous literature that addresses the competition between the players with the same function, such as various manufacturers making the homogenous products, and the complexity resulting from so many actors in the supply chain network system. By the concept of equilibrium, Nagurney et al. [4] explore in the general supply chain network setting. Other researchers expand the work of Nagurney et al. [4]. In particular, Dong et al. [5] study the supply chain network equilibrium with stochastic market demand which need get the density function or distribution function of random demand from history data.

In practice, demand uncertainties arise from the complexity and the evolvement over time of supply chain network which is actually a dynamic system [6]. The dynamics of our world results in the changing of network construct; thus we can discrete the fixed time into several planning periods, and in one planning period, the parameters in the network are stable, whereas in different periods, there are some changes such as the raw materials price fluctuation or the demand parameter transformation in the markets. In this paper, we model the discrete dynamic supply chain network equilibrium.

Moreover, in order to promote the product, firms usually use some marketing strategies such as advertising. Advertising is a common marketing activity and is widely used by enterprises. Local advertising, which focuses on the local market, is mainly accomplished by the cooperation between manufacturer and retailer [7]. Since the retailer is closer and familiar with the consumers, she may have an efficient local advertising channel, and the manufacturer may provide the retailer a part of money for local advertising purpose. Warner Brothers, a maker of corsets, issued the first co-op agreement in 1903 [7]. From then on, the use of co-op advertising spreads to other industrials such as grocery stores and fashion, and the automobile is the most common user of cooperative advertising today.

The advocating of advertisement could make consumers learn about the characters and related knowledge of the products provided by manufacturers and retailers, so more consumers will purchase this product, which result in the total market share increasing. If we consider the advertisement strategy in a dynamic decision context, then the relationship among different periods must be taken into consideration. For example, the advertising investment in the current period also has some effects in the next periods, and this effect will reduce over time. This paper incorporates the co-op advertising investment strategy in discrete dynamic decision-making environment, and the investment will be shared by manufacturers and retailers; the sharing ratio is determined by negotiation between the two tire players. As we see in the numerical examples, it is interesting to note that the value of ratio does not impact the equilibrium results. Since the advertising strategy is an option that is underutilized, enterprises are unsure about the economic performance of advertising investment.

To mitigate the ambiguity about advertising investment for decision makers, in the paper, we model the role of advertising investment in a supply chain network over time. Similar to literatures of supply chain network, we assume the players in the same tier such as all manufacturers compete in a noncooperative fashion and the players in different tiers such as manufacturers and retailers must cooperate in order to agree with each other in transaction price and amount. In the network, decision makers including manufacturers and retailers need to decide on the appreciation level of advertisement investment so that they sell more products to demand markets to maximize the profit. To simplify problem studied, we will illustrate this point through numerical examples and consider the investment levels as a constant instead of a decision variable.

This paper is organized as follows. Section 2 gives assumptions and notations. In Section 3, we model the optimal behaviors of various players in supply chain network. In turn, we establish the equilibrium model of the whole network. Section 4 provides solution algorithm for the model established, and in Section 5, we illustrate the effective and managerial insights by numerical examples. Finally, in Section 6, we conclude the paper.

2. Literature Review

Over the past decades, in the context of supply chain, advertising strategy has grown up and becomes an important research topic in operations research and management area. Cooperative advertising generally has five different meanings [23]. In our research, we employ the first one that is vertical cooperative advertising which is also the most common comprehension. The manufacturers offer to share a certain percentage of the downstream retailers’ advertising expenditures [24]. We also refer the readers to the work of [23] and the literature therein to get a general review about advertising. Based on the time dependence of parameters and decision variables, Lei et al. [25] and Xiao et al. [26] propose various multiperiod models to illustrate the impact of advertising investment on supply chain, whereas Chen [27], He et al. [28], Tsao and Sheen [29], and Xiao et al. [26] pick up the topic of stochastic environment associated with advertisement. Using game theoretic methods and from two main parts, simple marketing channels and a more complex structure, Jogensen and Zaccour [30] survey the literature on cooperative advertising in marketing channels (supply chains). Considering corporate social responsibility, Zhang et al. [31] examine the effectiveness of an advertising initiative in a leader-follower supply chain with one manufacturer and one retailer. Lambertini [32] characterizes an optimal two-part tariff specified as a linear function of the upstream firm’s advertising effort, performing this task both in the static and in the dynamic games. It is necessary to point out that these researches mainly pay attention to the simple supply chain or a single firm but do not consider the complexity and the mutual impacts among firms in the supply chain network.

Besides the research of Dong et al. [5], Nagurney et al. [8], Nagurney and Toyasaki [9], Wu et al. [10], Hammond and Beullens [11], Yang et al. [12], Masoumi et al. [13], and Yu and Nagurney [15], Toyasaki et al. [16] study the supply chain network equilibrium problems from various perspectives and different supply chain networks. Qiang et al. [14] establish a closed-loop supply chain network model considering the competition, distribution channel investment, and demand uncertainties. The literatures mentioned above deal with static supply chain or static supply chain network equilibrium problems.

Recently, a few authors explore supply chain network equilibrium problems in dynamic setting. For example, Cruz and Wakolbinger [17] develop a framework for the analysis of the optimal levels of corporate social responsibility (CSR) activities in a multiperiod supply chain network consisting of manufacturers, retailers, and consumers and describe the problem of carbon emissions. Daniele [18] considers a supply chain network model with three tiers of decision makers (manufacturers, retailers, and consumers) in the case when prices and shipments are evolving on time. Cruz and Liu [19] analyze the effects of levels of social relationship on a multiperiod supply chain network with multiple decision makers associated at different tiers. Hamdouch [20] establishes a three-tier equilibrium model with capacity constraints and retailers’ purchase strategy from a multiperiod perspective. Liu and Cruz [21] provide an analytical framework to investigate how financial risks affect the values of interconnected supply chain firms from a network perspective and how financial risks affect the supply chain firms’ profitability and the cash and credit transactions. Feng et al. [22] develop a closed-loop supply chain super network model in which the demand is seasonal and the manufacturers invest the reverse distribution channel for advocating consumers to return more end-of-life products.

The metamorphosis of supply chain network equilibrium literature of recent years is reviewed in Table 1. From Table 1 and literature survey, it is clearly evident that there is no research on discrete dynamic supply chain network equilibrium with advertising strategy and demand uncertainties.


No.AuthorsStatic/dynamicDemand characteristicConsidering factor

1Nagurney et al. (2002) [4]StaticDeterministic No
2Dong et al. (2004) [5]StaticRandom uncertaintyNo
3Nagurney et al. (2005) [8]StaticRandom uncertaintyB2B transaction, risk
4Nagurney and Toyasaki (2005) [9]StaticDeterministicNo
5Wu et al. (2006) [10]StaticDeterministicPollution tax
6Hammond and Beullens (2007) [11]StaticDeterministicCollection
7Yang et al. (2009) [12]StaticDeterministicCollection
8Masoumi et al. (2012) [13]StaticDeterministicNo
9Qiang et al. (2013) [14]StaticRandom uncertaintyChannel investment
10Yu and Nagurney (2013) [15]StaticDeterministicNo
11Toyasaki et al. (2014) [16]StaticDeterministicNo
12Cruz and Wakolbinger (2008) [17]Discrete dynamicDeterministicCorporate social responsibility
13Daniele (2010) [18]Continuous dynamicDeterministicNo
14Cruz and Liu (2011) [19]Discrete dynamicDeterministicSocial relationship
15Hamdouch (2011) [20]Discrete dynamicDeterministicPurchase strategy
16Liu and Cruz (2012) [21]Discrete dynamicDeterministicCorporate financial risks, trade credits
17Feng et al. (2014) [22]Continuous dynamicDeterministicChannel investment

In this paper, our model captures the planning process and the change of costs and demands and highlights the performance of advertising with delay effect, and moreover, this model expresses the uncertainties popularly existing in practice.

3. Model Assumptions and Notations

3.1. Model Assumptions

We consider a supply chain network consisting of suppliers, manufacturers, retailers, and demand markets and let denote a typical supplier, a typical manufacturer, a typical retailer, and a typical demand market; a retailer is matching a demand market; that is, one retailer only deals with the demand of one demand market. All actors in the same tire compete in a noncooperative fashion. Figure 1 illustrates the simple supply chain network with 2 suppliers, 2 manufacturers, 2 retailers, and 2 demand markets in 2 periods. denotes the first supplier in the first period, and denotes the second supplier in the first period; the other notations can be explained in the same way. The real lines between two adjacent tiers denote the related transaction activities, and the dash lines between 2 periods denote inventory transferring from the former period to the latter period.

In order to explicate the problem studied, we give the following assumptions:(1)All vectors are column vectors;(2)The equilibrium solution or the optimal value of a decision variable is denoted by “*”;(3)The advertising investment is a constant and shared between the pairs of manufacturer and retailer;(4)All cost functions and transaction functions are continuous convex and differentiable;(5)All players in the network are risk neutral.

3.2. Variables and Notations

The variables and notations are defined as in Tables 2 and 3, and the production functions and transaction functions are defined as in Table 4.


NotationDefinition

Raw material conversion rate
A typical period,
A typical supplier,
A typical manufacturer,
A typical retailer,
A typical demand market,
>0The unit cost of product shortage of retailer
>0The unit cost of product excess of retailer
Advertising investment ratio shared by manufacturer


NotationDefinition

The raw material transaction volume from supplier to manufacturer at period ; group all of these variables into a column vector
The total raw material volume provided by supplier to all manufacturers at period ; group all of these variables into a column vector
The total raw material volume of manufacturer used to produce at period ; group all of these variables into a column vector
The product transaction volume from manufacturer to retailer at period ; group all of these variables into a column vector
The transaction price charged by manufacturer for retailer at period
Price charged by retailer to the product in his outlet for corresponding demand market at period
The density function of random variable
The distribution function of random variable
The inventory of manufacturer at period ; group all of into a column vector


NotationsDefinition

The raw material production cost function of supplier at period
The transaction cost function between supplier and manufacturer at period
The production cost function using raw materials of manufacturer at period
The exhibition and disposal cost at retailer at period
The transaction cost function between manufacturer and retailer at period
The inventory cost function at manufacturer
The delay effect factor of advertising investment at period on the period
The demand function associated with demand market

4. Discrete Dynamic Supply Chain Network Equilibrium Model

4.1. The Optimal Behavior and Equilibrium Condition of Suppliers

In each period, supplier provides raw material to various manufacturers at the beginning of every period and makes decision associated with trade and production volume of raw material to maximize the profit in the entire planning horizon. Using the notations defined previously, the profit maximum criterion for supplier can be described as

Equation (2) expresses that production output of raw material cannot be lower than total volume of the raw material transaction between the supplier and the various manufacturers.

In this paper, we assume that all the suppliers compete in a noncooperative fashion. Therefore, we can simultaneously express the equilibrium condition of the suppliers as the variational inequality, determining , such that where .

In (4), is the Lagrange multiplier corresponding to constraint (2) and is the column vector with the elements of .

Based on the equivalence of variational inequality and complement problem, from the second term of (4), we get

From the 1st term of (4), in the equilibrium state, ; that is, is equal to the marginal production cost. Therefore, (5) shows that the transaction price between suppliers and manufacturers is equal to the sum of marginal transaction cost and marginal production cost.

4.2. The Optimal Behavior and Equilibrium Condition of Manufacturers

The manufacturers purchase the raw materials from various suppliers to make products and sell the new products to retailers at every period and in the same time manage inventory between periods according to the market conditions. The manufacturer seeks to maximize her profit that can be described as follows:

Equation (7) expresses the flow conservation; the sum of production volume from raw materials in period and the transferring inventory from period is equal to the sum of the transaction volume with all retailers and the transferring inventory to next period, and assume the corresponding Lagrange multiplier is ; is the column vector with the elements of . Equation (8) shows that the raw materials amount obtained in manufacturer is not higher than that various suppliers sent to her; similarly, assume the corresponding Lagrange multiplier is and is the column vector with the elements of .

The profit maximum object of all manufacturers can be described as a variational inequality, determining , such that where .

From the third term of (9), the transaction price can be written as when the network is in equilibrium:

From the 2nd term of (9), in the equilibrium state, we get ; then from the 1st term, we get . Equation (10) shows that in the equilibrium state, the transaction price between manufacturers and retailers is equal to the sum of marginal transaction cost between manufacturers and retailers, the Lagrange multiplier corresponding to constraint (7), and the advertisement investment amount shared by manufacturer .

4.3. The Optimal Behavior and Equilibrium Condition of Retailers

The retailers need to decide to purchase how many products from manufacturers and sell to consumers in corresponding demand markets in a certain price.

Due to denoting the random demand of retailer outlet , the demand depends on the advertising investment and the trade price; it is obvious that the more advertising investment paid by manufacturers and retailers is, the larger consumer demand is, whereas the increase of price charged by retailers will lower the product demand. For a given product transaction price at period , according to the notations illustrated in Table 3, . Let denote the wholesale amount from manufacturers and ; group all in period into a column vector , and group all into a column vector . In order to express the competition among retailers, we assume that the exhibition function and disposal cost function at retailer are related with all retailers.

For retailer , if given , it is similar as in Dong et al. [5] and Nagurney et al. [4], the expected sales quantity, expected shortage quantity, and expected exceed quantity can be expressed as

From (10), we can easily obtain

For retailer , the maximum expected profit model can be expressed as

Using (11) and (13) can be rewritten as

All retailers compete in a noncooperation fashion; using (12), their equilibrium conditions can be described as a variational inequality, determining , such that where .

In (16), is the Lagrange multiplier corresponding to constraint (14) and is the column vector with the elements of . The transaction price is a decision variable which can be obtained from the computing results.

4.4. The Optimal Behavior and Equilibrium Condition of Demand Markets

For the supply chain network, given a fixed advertising investment, the consumers of demand markets buy the products under a price charged by the retailers and it is similar as in Dong et al. [5] and Nagurney et al. [4]

The consumers’ optimal behaviors and equilibrium conditions can be described as a variational inequality, determining , such that where .

4.5. The Equilibrium Condition of the Supply Chain Network

Each player in the supply chain network selects the optimal strategy in every period and seeks to maximize the profit in the entire planning horizon on the basis of the other players making optimal decisions. Thus, the network will experience a strategy selecting process and carry out Nash equilibrium in the end. In particular, the product transaction amount and price between the adjacent tires must be equal to that the players want to purchase or sell at every period, and the manufacturers and retailers also need to make decisions about the advertising investment to enhance the expected sales to maximize their profits. So, the whole network equilibrium condition is the sum of (4), (9), (15), and (18). We sum up these equations and obtain the following theorem.

Theorem 1. A strategy pattern of the discrete dynamic supply chain network can be called an equilibrium pattern if and only if it satisfies the following inequality, determining , such that where .

Proof. Let us sum up (4), (9), (15), and (18); we get the total inequality, determining , such that We simplify the 3rd and 4th terms in (20) and obtain (19). From (19), we note that the share ratio of advertising investment between manufacturers and retailers does not impact the network equilibrium results; therefore, determining the share ratio will be up to the power of two kinds of players in their bargain.

5. Numerical Examples

In this section, we will provide some numerical examples to illustrate the efficiency of the previous equilibrium model and analyze the relevant parameters. To solve the model, there are several algorithms to choose, such as logarithmic-quadratic proximal prediction-correction method [33], modified contraction project method [34], smoothing Newton algorithm [35], and others, to name a few. In this paper, we employ the modified contraction project method to solve the variational inequality (19) for its simple steps and obtain the decision variables and Lagrange multiplexer simultaneously. Set the related parameters as follows: the initial value of decision variables and Lagrange multipliers is set to 1 and the convergence criterion, for example, the absolute value of difference of decision variables and Lagrange multipliers between two steps is lower than or equal to 10−8. We assume , , , , , , , and . The related cost functions and parameters are set as listed in Table 5. It is assumed that the random demands follow uniform distribution in, , , , and , for , , , and .


NotationDefinition

Cost function of producing raw materials for supplier at period
Transaction cost function undertaken by supplier related to supply chain at period
Production cost function for manufacturer at period
Inventory cost for manufacturer at period
Transaction cost function undertaken by manufacturer related to supply chain at period
Disposal costs at retailer at period

This paper focuses on the analysis of the following four aspects: (1) the equilibrium results of advertising investment with delay effect and the results listed as in Table 6; (2) the equilibrium results of advertising investment with no delay effect, that is, , and the results listed as in Table 7; (3) the equilibrium results with one manufacturer advertising investment and the results listed as in Table 8; and (4) the profits of various players with the 1st period advertising investment increasing with/without delay effect, which is illustrated as in Figure 2.


Variables 



0.72820.72590.72140.74070.73400.7488
0.44430.44310.44080.4510.44740.4552
0.38210.38130.37970.38670.38420.3896

,

1.45631.45191.44291.48141.46791.4976
0.88860.88630.88150.90190.89480.9105
0.76430.76270.75940.77330.76850.7791



0.53570.51740.51820.53360.51610.5322
0.52280.53130.51320.53990.53000.5386
0.49610.50170.51060.50480.51950.5228


0.38490.41700.40660.41420.43580.4331
0.22800.24070.26170.23640.27060.2664
000000

3.32633.29283.22543.51703.41433.6415
14.323414.242914.084514.757214.516315.0348
27.200427.071626.772128.167427.735528.8402


Variables