PETRI NET MODELLING FOR SUPPLY CHAIN OF PERISHABLE PRODUCTS

Abstract

Decentralized supply chains results into different types of inefficiencies particularly in case of perishable product supply chains (PPSC). The thesis is an attempt to use Petri Net Modelling to improve the performance of PPSCs. The first objective of the thesis focused on assessing failures associated with PPSC for evaluating the performance in terms of effective PPSC management adoption using fault tree methodology. Different failure events were identified from literature and semi-structured interview from experts. Fault tree model was developed from the identified failure events. Probability of failure events were calculated using Poisson distribution based on the annual reports and interviews conducted from experts. In this study, fault tree development and analysis were conducted on apple supply chain to present the authenticity of this method for failure analysis. The findings indicate that the failure events at production and procurement, involvement of middleman, handling and packaging failure, and transportation failure hold the highest-ranking scores. Further the study focused on establishment of localized processing unit for long-term project as strategic decision making in supply chain. Optimal location of processing unit is of great importance where high quantities of agricultural production is spread over fragmented farm size with large number of farmers. In our study we proposed two techniques: Mixed Integer Linear Programming (MILP) and Colored Petri Nets (CPN) to optimize and analyse processing unit and partner selection problem under uncertainty to reduce cost and enhance efficiency under uncertain circumstances. The mathematical model was set up in regard with capacity of processing unit, transportation cost, and partner selection. The optimized model discusses production-planning including strategic and operational decisions under disrupted environment. The model simulated in concurrent effects of disruption categorized in terms of probability associated with economic terms. Thereby, the model can be used as a decision-making system for processing companies to result in efficient PPSC management. The developed model was tested and validated on case study to demonstrate our approach’s applicability. Though only one case study of apple supply chain data is used to demonstrate the results but it can be used for similar PPSCs. As India has multiple agri-product based supply chains which need more efficient solutions, present study opens a new research area using Petri Net Modelling for handling these PPSCs.