Welcome!
I am a PhD scholar (expected to graduate in 2026) at the Indian Institute of Technology, Goa, working under the guidance of Dr. Neha Karanjkar.
My research focuses on the simulation modeling, analysis, and optimization of supply chains.
I am currently exploring Graph Neural Network (GNN) metamodels for simulation-based optimization of supply chains. A metamodel is a coarser model that approximates the input-output behavior of a simulation as a smooth surface, so that it can stand in for the simulation inside an optimization loop. Supply chains have a natural graph structure, and a GNN metamodel exploits it: the metamodel captures the dependencies between the nodes and links of a network and generalizes across topologies. A single GNN metamodel can therefore be trained on supply chain networks of arbitrary structure and complexity. GNNs are also differentiable, which opens the prospect of combining structural optimization (network topology) with parametric optimization (replenishment policies).
The training data for the metamodels comes from SupplyNetPy, an open-source Python library that we developed for modeling and simulating arbitrary supply chain and inventory networks.
Research Interests: Discrete Event Simulation, Simulation-based Optimization, Supply Chains, Operations Research, Digital Twins, and Graph Neural Networks.
Thank you for stopping by!
I am currently exploring Graph Neural Network (GNN) metamodels for simulation-based optimization of supply chains. A metamodel is a coarser model that approximates the input-output behavior of a simulation as a smooth surface, so that it can stand in for the simulation inside an optimization loop. Supply chains have a natural graph structure, and a GNN metamodel exploits it: the metamodel captures the dependencies between the nodes and links of a network and generalizes across topologies. A single GNN metamodel can therefore be trained on supply chain networks of arbitrary structure and complexity. GNNs are also differentiable, which opens the prospect of combining structural optimization (network topology) with parametric optimization (replenishment policies).
The training data for the metamodels comes from SupplyNetPy, an open-source Python library that we developed for modeling and simulating arbitrary supply chain and inventory networks.
Research Interests: Discrete Event Simulation, Simulation-based Optimization, Supply Chains, Operations Research, Digital Twins, and Graph Neural Networks.
Thank you for stopping by!
News
- 2026 Two papers accepted at the Winter Simulation Conference (WSC 2026): the SupplyNetPy simulation library, and a study on GNN metamodels for supply chain optimization.
- Jun 2024 Presented “Development of an open-source library for supply chain modeling and optimization” at ACM SIGSIM-PADS 2024 (online); awarded an ACM travel grant.
- Dec 2023 Presented an InventOpt poster at IndoML 2023, IIT Bombay.
- Jul 2023 Presented “An Open Tool-Set for Simulation, Design-Space Exploration and Optimization of Supply Chains and Inventory Problems” at SIMULTECH 2023, Rome (online).
Tushar Lone
PhD Scholar,School of Mathematics and Computer Science,
Indian Institute of Technology, Goa
tusharlone28{at}gmail{dot}com
tushar20231102{at}iitgoa{dot}ac{dot}in
Download CV (PDF)
