Predicting Sustainable Supply Chain Performance Based on GRI Metrics and Multilayer Perceptron Neural Networks

The sustainable supply chain performance prediction model anticipates the values of key performance indicators also known lagging metric, based on the leading metrics. The prediction system helps decision-makers to identify the performance gaps and helps to take necessary action plans to minimize the deviation between the targets set and the outputs that are estimated by the model. This study uses a worldwide accepted Global Reporting Initiative (GRI) metrics and evaluates values of level 1 metrics that are effects of level 2 metrics by artificial neural networks(ANN). The literature presents several sustainable supply chain performance evaluation models; however, a prediction model based on a combination of GRI and ANN is a fairly unexplored area. Multilayer perceptron neural network model has the ability to adjust with the environment of use with the help of past performance data unlike models present in the literature that require manual parameterization for updating and implementing them. MATLAB is used for the computational implementation of ANN models. From the results of Pearson correlation coefficient, the correlation obtained among the targeted and forecasted performance values for the GRI level 1 metrics is positive.