Search results

The purpose of this paper is to describe a real‐world system developed for a large food distribution company which requires forecasting demand for thousands of products across multiple warehouses. The number of different time series that the system must model and predict is on the order of 10⁵. The study details the system's forecasting algorithm which efficiently handles several difficult requirements including the prediction of multiple time series, the need for a continuously self‐updating model, and the desire to automatically identify and analyze various time series characteristics such as seasonal spikes and unprecedented events.

The forecasting algorithm makes use of a hybrid model consisting of both statistical and heuristic techniques to fulfill these requirements and to satisfy a variety of business constraints/rules related to over‐ and under‐stocking.

The robustness of the system has been proven by its heavy and sustained use since being adopted in November 2009 by a company that serves 91 percent of the combined populations of Australia and New Zealand.

This paper provides a case study of a real‐world system that employs a novel hybrid model to forecast multiple time series in a non‐static environment. The value of the model lies in its ability to accurately capture and forecast a very large and constantly changing portfolio of time series efficiently and without human intervention.

The purpose of this paper and its companion (Part II: multi‐silo supply chains) is to investigate methods to tackle complexities, constraints (including time‐varying constraints) and other challenges. In tis part, the paper aims to devote attention to single silo and two‐silo supply chains. It also aims to discuss three models. The first model is based on the winebottling real‐world system and exposes complexities of a single operational component of the supply chain. The second model extends it to two components: production and distribution. The last system is a real‐world implementation of the two‐component supply chain.

Evolutionary approach is proposed for a single component problem. The two‐component experimental supply chain is addressed by the algorithm based on cooperative coevolution. The final problem of steel sheet production is tackled with the evolutionary algorithm.

The proposed systems produce solutions better than solutions proposed by human experts and in a much shorter time.

The paper discusses various algorithms to provide the decision support for the real‐world problems. The proposed systems are in the production use.

The purpose of this paper and its companion (Part I: single and two‐component supply chains) is to investigate methods to tackle complexities, constraints (including time‐varying constraints) and other challenges. In this part, attention is devoted to multi‐silo supply chain and the relationships between the components. The first part of the paper aims to consider two types of experimental supply chains: with one‐to‐many and many‐to‐one relationships. The second half of the paper aims to present two approaches on optimising the material flow in the real‐world supply chain network.

Cooperative coevolutionary and classical sequential approaches are taken to address the experimental multi‐silo supply chains. Due to the nature and the complexity of the supply chain presented in the second half of the paper, evolutionary algorithm was not sufficient to tackle the problem. A fuzzy‐evolutionary algorithm is proposed to address the problem.

The proposed systems produce solutions better than solutions proposed by human experts and in much shorter time.

The paper discusses various algorithms to provide the decision support for the real‐world problems. The system proposed for the real‐world supply chain is in the process of integration to the production environment.