Ilya Jackson

Massachusetts Institute of Technology | MIT · Center for Transportation and Logistics

Considering recent success of neuroevolutionary approaches in automated machine learning, their application to metamodeling seems to be promising. This paper examines, whether it is feasible and efficient to combine artificial neural network with genetic algorithm for metamodeling automation of logistic and production systems. Besides, the possibil...

Adaptive and highly synchronized supply chains can avoid a cascading rise-and-fall inventory dynamic and mitigate ripple effects caused by operational failures. This paper aims to demonstrate how a deep reinforcement learning agent based on the proximal policy optimization algorithm can synchronize inbound and outbound flows and support business co...

Decision-making in supply chains is challenged by high complexity, a combination of continuous and discrete processes, integrated and interdependent operations, dynamics, and adaptability. The rapidly increasing data availability, computing power and intelligent algorithms unveil new potentials in adaptive data-driven decision-making. Reinforcement...

A typical retailer carries 10,000 stock-keeping units (SKUs). However, these numbers may exceed hundreds of millions for giants such as Walmart and Amazon. Besides the volume, SKU data can also be high-dimensional, which means that SKUs can be segmented on the basis of various attributes. Given the data volumes and the multitude of potentially impo...

Synchromodality is an emerging concept in supply chain management. A synchromodal supply chain can be defined as a multimodal transportation planning system, wherein the different agents work in an integrated and flexible way that enables them to dynamically adapt the transport mode based on real-time information from stakeholders, customers, and t...

Understanding the structure and dynamics of the supply chain enables businesses to optimize their production and distribution flows and to prepare for potential disruptions.

As one of the world’s largest logistics platforms, C.H. Robinson solves logistics problems for companies across the globe and across industries, from the simple to the most complex. Motivation: Inventory management is a crucial concept in the world’s increasingly complex and global supply chain. Poor inventory management not only leads to suboptima...

The paper discusses the prospects for the development and implementation of centralized ground traffic control systems at airports. The automatic control system can only work if there is accurate data on the location of mobile objects, which include both vehicles involved in the maintenance of aircraft and the aircraft themselves. In order to devel...

Our work is the first attempt to apply Natural Language Processing to automate the development of simulation models of logistics systems. We demonstrated that the framework built on top of the fine-tuned Transdormer-based language model could produce functionally valid simulations of queuing and inventory control systems given the verbal descriptio...

Synchronized supply chains can mitigate a cascading rise-and-fall inventory dynamic and prevent cycles of over and under-production. This paper demonstrated that a deep reinforcement learning agent could only perform adaptive coordination along the whole supply chain if end-to-end information transparency is ensured. Operational and strategic disru...

In the research, the aspects of decision-making according to the airport activities were considered. The decision about airport planning and management should be comprehensive and operative and of course, the assessment of alternative decisions is necessary. The purpose of this research is to highlight the role of simulation modelling at the stage...

Supply chain synchronization can prevent the "bullwhip effect" and significantly mitigate ripple effects caused by operational failures. This paper demonstrates how deep reinforcement learning agents based on the proximal policy optimization algorithm can synchronize inbound and outbound flows if end-to-end visibility is provided. The paper conclud...

Floating car data (FCD) have recently become a popular tool of urban traffic engineering. Conventional FCD contains a series of probe cars’ timestamped locations and are used to estimate traffic speeds and travel times and identify congestions. In this study, we propose the enhancement of conventional FCD with car vision information: traffic measur...

The paper describes the experience of development and applications of a simulation program GTSS. The program is designed to develop and test algorithms for centralized control of ground transport processes at airports. The models built using GTSS take advantage on the discrete time counting principle “delta T” and the discretization of the 2D airpo...

Markov-modulated linear regression (MMLR) model is a special case of Markov-additive processes. The model assumes that unknown regression coefficients depend on an external state of the environment, but regressors remain constant. MMLR model differs from other switching models by a new analytic approach to parameter estimation and known transition...

The paper describes experiments with a Digital Twin, which is intended to be used as a testbed for solutions in the field of centralized traffic control at airports. The data flow on spatial characteristics of vehicles is simulated using a special simulation model. A distinguishing feature of the model is the ability to describe and reproduce speci...

Inventory control has been a major point of discussion in industrial engineering and operations research for over 100 years. Various advanced numerical methods can be used for solving inventory control problems, which makes it a highly multidisciplinary filed attracting researchers from different academic disciplines. This fact makes it a daunting...

Taking into consideration the urgent industrial need in metamodeling automation and recent success of neuroevolutionary approaches in neural architecture search and hyperparameter optimization, this thesis examines feasibility and efficiency of the combination of artificial neural network and genetic algorithm for automated metamodeling of inventor...

This paper discusses the application of neuroevolutionary automated machine learning to metamodeling of complex production-inventory systems. The proposed framework incorporates multilayer perceptron and genetic algorithm. As a numerical example this paper also demonstrates the application of this framework to metamodeling of multiproduct productio...

Because of seasonality of demand and periodicity of replenishments, inventory dynamics can be highly self-similar. This paper demonstrates that such metrics of fractal analysis as the Hurst exponent, correlation dimension and sample entropy indicate predictability of inventory dynamics by LSTM recurrent neural networks. From business point of view,...

In order to tailor inventory control to urgent needs of grocery retail, the discrete-event simulation model with realistic perishability mechanics is proposed. The model is stochastic and operates with multiple products under constrained total inventory capacity. Besides that, the model under consideration is distinguished by quantity discount, unc...

Inventory control problems arise in various industries, and each single real-world inventory is replete with non-standard factors and subtleties. Practical stochastic inventory control problems are often analytically intractable, because of their complexity. In this regard, simulation-optimization is becoming more and more popular tool for solving...

Inventory control problems arise in various industries, and each single real-world inventory is replete with non-standard factors and subtleties. Practical stochastic inventory control problems are often analytically intractable, because of their complexity. In this regard, simulation-optimization is becoming more and more popular tool for solving...

This study was conducted within the course “Data mining”, which was taught during the fall semester 2018 at the Transport and Telecommunication Institute by Dr. Onieva. This study analyzes Latvian household budget and consumption in 2012. The data is based on survey conducted by the Central Statistical Bureau of Latvia. This study is mainly focused...

Baseline data

The presentation demonstrated during the Winter Simulation Conference 2018

This paper proposes a discrete-event model of a multi-product (Q, r) inventory control system. The described approach provides a computationally efficient method to estimate a current inventory policy and test alternatives. In the considered model, values of a reorder level r and a reorder quantity Q are approached through iterative methods. The mo...

This paper reports on the unsupervised learning approach for solving stock keeping units segmentation problem. The dataset under consideration contains 2279 observations with 9 features. Since the “ground truth” is not known, the research aims to compare such clustering algorithms as K-means, mean-shift and DBSCAN based only on the internal evaluat...

This paper describes the design and application of a model simulating the passenger flows on the high-speed rail line Rail Baltica. The proposed model allows one to define passenger flow service indicators, according to data on the behavior of potential passengers that take the train in different localities along Rail Baltica. If corresponding base...

The paper describes an eventual combination of discrete-event simulation and genetic algorithm to define the optimal inventory policy in stochastic multi-product inventory systems. The discrete-event model under consideration corresponds to the just-in-time inventory control system with a flexible reorder point. The system operates under stochastic...