Elsevier

Journal of Cleaner Production

Volume 231, 10 September 2019, Pages 1136-1148
Journal of Cleaner Production

Evolutionary framework design in formulation of decision support models for production emissions and net profit of firm: Implications on environmental concerns of supply chains

https://doi.org/10.1016/j.jclepro.2019.05.300Get rights and content

Abstract

There have been increased investments in cleaner technologies and adoption of a voluntarily limit on transportation emissions by the global firms to handle the environmental concerns of supply chains and to increase demand for finished goods. Consequences are the reduction in net profit for the firm. To address this trade-off between the net profit and environmental concerns, the formulation and optimization of a compact model are needed. Development of these models requires a thorough understanding of the nature of the impact of three inputs (investment coefficient, penalty per unit emission and customer's emission elasticity) on production emissions and net profit. Past studies revealed that a compact model comprising the interactive effect of these inputs on the production emissions and net profit is not yet formulated. Therefore, this study illustrates the development of an evolutionary framework of an advanced multi-gene genetic programming in the formulation of functional expressions for the net profit and production emissions based on the three inputs (investment coefficient, penalty per unit emission and customer's emission elasticity) of the monopolist firm. The sensitivity and parametric based 2-D analysis determine the relationships and found that the penalty per unit emission is dominant input for reducing emissions and maintaining net profit simultaneously. The contribution of this work lies in designing of an evolutionary framework in the development of empirical explicit expressions, which can easily be optimized analytically to keep production emissions and net profit balanced.

Introduction

With global climate deteriorating over the years, it has alerted the governments to initiate the campaigns for cleaner and safer earth and impose stringent regulatory norms on industries in monitoring the carbon emissions. The market has become more considerable about the environment due to widespread awareness related to green technology, hence gaining market value by producing environmentally friendly products (Jacobs and Subramanian, 2012). As of late, ecological concerns are raised by developing nations against the developed nations for outsourcing the contaminating assembling units. In the United States, it is discovered that around 75% of clients buy items remembering the environmental consciousness of the organization (Carter et al., 2000) and 80% of them will pay more for eco-friendly items (Harris, 2006). This demonstrates that environmental performance alongside productivity has turned into an imperative trademark for judging the notoriety of any organization. Henceforth, this has forced industries to get acquainted with and execute greener practices to ensure the production and distribution of eco-friendly products.

Easy correspondence and reduction in transportation cost enable products to be delivered to remote locations which open up avenues for the global market (Lam and Dai, 2015). Moreover, increased global transportation of products has significantly heightened up the carbon emissions. This has additionally prompted an expansion in off-shoring practices and decentralization of manufacturing processes. It was recommended that if these emanations can be incorporated as a tariff in total production expenditure (Lam and Dai, 2015; Choudhary et al., 2015), the manufacturers might feel coerced to provide an environment-friendly label to products. This label will depict how much greener their manufacturing practices are and how much distance their product has traveled (Sundarakani et al., 2010; Hua et al., 2011). Enterprises face two difficulties: 1) significant reduction in demand 2) penalty by regulators if their manufacturing practices and environment are not greener. (Sarkis et al., 2011; Kleindorfer et al., 2005). Experts have indicated that the financial health of the company can also be affected even if the carbon emissions have not occurred within a company but within the supply chain of the company (Busch and Hoffmann, 2007; Kassinis and Soteriou, 2003). In brief, past studies have broadly discussed following three main inputs/factors that address the environmental aspects in supply chains of a monopolistic firm (Choudhary et al., 2015):

This factor measures the evaluation of the customer's attitude towards eco-friendly products. Higher the values of emission elasticity, more the customers are conscious of the environmental aspect of the product (Klassen and McLaughlin, 1996; Yalabik and Fairchild, 2011). Even in an increasingly environmentally conscious market, manufacturing firms and consumers are not ready to share equal liability towards environmental repercussions. (Weber and Matthews, 2008; Gay and Proops, 1993). Two widely used principles have been adopted widely: 1) Consumer responsibility principle which states that the consumer that consumes goods is responsible for environmental consequences rather than the production units and 2) Producer responsibility principle delineates that the place or nation where the manufacturing is carried out is responsible for environment (IPCC, 1997). There are hardly any studies exists that take into account both the principles (Choudhary et al., 2015).

Greener technology development in industry and practice of policies do have resulted in curtailing of carbon emissions (Hart, 1995). However, the effect of these investments is only for a shorter term but in long more sustainable and evolvable policies and framework is needed (Popp, 2005). Such investments may diminish the net benefit of the firm in a long run. Researchers have carried out life cycle analysis (Yalabik and Fairchild, 2011), suggested optimal green product design (Chen and Sheu, 2009), and formulated new production capacity planning models (Angell and Klassen, 1999).

With transportation reaching remote locations and government adopting liberal trade policies with other countries has resulted in the creation of global supply chains (Nijkamp, 1999). Finished goods are delivered to remote areas by different modes of transport, for example, rail, road, ocean, and water through third-party logistics providers (Cholette and Venkat, 2009). The firm may adopt a voluntary limit on its emissions from transportation. However, such investments on greener means of transportation and impositions (voluntarily limit/carbon emission tax) considered can also reduce the net benefit of the firm (Choudhary et al., 2015). The Network for Transport and Environment (NTM) method (Lam, 2015) based on the selection of transportation mode with self-imposed emission limit was commonly used to calculate carbon emissions from different methods. In this study, the carbon/production emissions correspond to the fuel consumption of operations in industry and the emissions during transportation of goods from the third party logistic provider.

The works described above indicated that the factors relating to environmental concerns in supply chains were tackled individually but hardly any studies exist that considers all factors comprehensively to provide a model or solution (Fig. 1) (Choudhary et al., 2015). Recently, a holistic model based on non-linear programming (NLP) has been developed by Choudhary et al. (2015) which propounded a trade-off between net benefit and production emissions for a monopolist firm. Tiwari (Tiwari et al., 2015) in their editorial work focused on the requirement for investigating the utilization of heuristic strategies in devising decision support models for carbon emissions and net benefit. This work also highlights the dearth of research works in the field of decision support models (Zheng and Wang, 2014; Choudhary et al., 2014) to address the environmental concerns in supply chains (see Fig. 2).

In the context of the formulation of decision support models, the statistical methods such as response surface methodology (RSM) can be implemented. The main problem with RSM is that it cannot be used for the data beyond the training range. In the circumstance of uncertain system behavior, the non-conventional methods such as genetic programming (GP), artificial neural network and support vector regression seems a suitable alternative. Among these methods, GP has the ability to evolve the explicit models (functional relationships) without the need for incorporation of expert knowledge (Hinchliffe et al., 1996). Significant work in the development of variants of GP has been made. A most recent variant of GP, multi-gene genetic programming (MGGP) has received greater attention due to its ability to evolve models from the multiple sets of genes (Hinchliffe et al., 1996; Gandomi and Amir Hossein Alavi, 2012). These multiple genes combining the ability of the MGGP also increases its vulnerability towards the formation of complex (over-sized) models which over-fit on the data. Based on preliminary analysis on MGGP, three shortcomings, namely (1) inappropriate practice of formulation of the MGGP model, (2) inappropriate complexity measure of the MGGP model, (3) complexities in model selection are identified (Garg and Lam, 2015).

Therefore, this study illustrates the design of evolutionary based advanced multi-gene genetic programming (AMGGP) framework in the formulation of functional expressions for production emissions and net benefit of a monopolistic firm based on the three inputs i.e., investment coefficient, penalty per unit emission and customer's emission elasticity. The designed framework will address the three shortcomings in the framework of MGGP and will be beneficial for practitioners to find the optimum balance of net profit and production emissions without much need of human intervention. Its performance is then evaluated against the actual data based on cross-validation, error metrics and goodness-of-fit tests. Further, the nature of the impact of these inputs on the production emissions and net profit is analyzed by 2-D analysis of the formulated models. The sensitivity and parametric based 2-D analysis determine the relationships and found that the penalty per unit emission is dominant input for reducing emissions and maintaining net profit simultaneously. The contribution of this work lies in designing of an evolutionary framework in the development of empirical explicit expressions, which can easily be optimized analytically to keep production emissions and net profit balanced.

Section snippets

Research problem and data acquisition

Generating the optimum balance of net profit with decreased production emission the main problem (Choudhary et al., 2015). An increment in speculations for greener innovation incites the ecological cognizant clients and creates a more prominent interest for environmentally friendly products. Nonetheless, the ventures and volunteer point of confinement on emanation in longer run may lessen the net benefit of the firm. As talked about in Section 1, it is imperative to comprehend the idea of the

Design of evolutionary framework of advanced multi-gene genetic programming

The principle of “Survival of the fittest” of Darwinian forms the working principle of genetic programming (GP) (Koza, 1992). The principle of GP is same as genetic algorithms (GA), but the difference between them lies in the representation of solutions. In GP, the solutions are represented by tree/structure representing a mathematical equation, whereas in GA the solutions are represented by a crisp value of inputs. Past studies reveal that significant development has been made in developing

K-fold cross-validation method to test the robustness of performance comparison of three models

K-fold cross-validation approach with k value of 5 is used to divide the data set as discussed in Section 2 into five random training and testing data sets. This is mainly done because the training data affects the learning ability of the algorithm and therefore randomization of the training data may also affect the generalization ability. The data sets partitioned into five is designated by Data set 1, Data set 2, Data set 3 … Data set 5 as represented in Table 3. The part of research data is

Conclusions

The present work has made a contribution by developing decision support models for the purpose of optimizing production emissions and net profit simultaneously. Two important contributions arising from the research work are 1. Development of new evolutionary framework based on AMGGP that addresses the issues in MGGP. 2. Formulation of decision support models for net profit and production emissions. The main inputs (three), namely penalty per unit emission, customer's emission elasticity and

Acknowledgement

Authors acknowledge Grant DMETKF2018019 by State Key Lab of Digital Manufacturing Equipment and Technology (Huazhong University of Science and Technology). Authors also like to acknowledge Shantou University Youth Innovation Talent Project (2016KQNCX053) supported by Department of Education of Guangdong Province and Shantou University Scientific research fund (NTF 16002).

References (36)

Cited by (0)

View full text