Modeling and analysis of demand for commodities and a case study of the petrochemical market (Q2873483)

This thesis deals with the development of a quantitative nonlinear demand model that reflects the influence of all products' prices in the market. The first chapter is an introduction. In the second chapter, an overview of existing demand modeling approaches is provided. The author presents models for demand of households based on the concept of neoclassical utility theory. Also, some issues with respect to demand are discussed that come from neoclassical production theory. The author reviews approaches to modeling the influence of substitution on demand in the context of assortment planning problems and describes various approaches to modeling demand for specific services or products. In the third chapter, a general description of the characteristics of a demand model for commodities is presented. In addition to the analysis of the relationship between demand and price, the author considers the effects of economic indices quantifying the change in the economic situation. The effect of the specific consumer's characteristics and the influences of other products' prices are studied. In the fourth chapter, there is presented a supply-demand trade network optimization model to simulate prices and sales quantities in a multi-commodity market. A review of existing production and sales optimization models in the field of supply chain management or chemical engineering is provided. Then the author proposes the approach to determine the optimal pricing, production, and transport strategies by maximizing the overall profit in a multi-regional and multi-commodity market. The fifth chapter deals with the theory of constrained optimization as well as algorithms to solve inequality constrained nonlinear optimization problems that are used throughout the thesis. Algorithms are described to solve least-squares problems such as the generalized Gauss-Newton method. In the sixth chapter, an explicit demand model is developed that can be integrated in the supply-demand trade network optimization model applied to a product network of the petrochemical industry. The author extends the basic model that reflects the relationship between demand and price, and includes model components for the influence of changes in the economic situation, as well as the influence of substitutable and complementary products. In the seventh chapter, methods to identify the demand parameters in the situation of incomplete data are investigated. The author proposes two different approaches. First, the heuristic method determines the parameters based on price and sales quantities simulations. Second, a constrained parameter identification problem is established that can be solved by the generalized Gauss-Newton method. The eights chapter presents the conclusion.