Industrial plants investment projects efficiency estimation based on simulation and artificial intelligence methods

Abstract

Methods and software enabling the estimation of efficiency and the comparisons of alternative designs of industrial plants are discussed. Problem of efficiency estimation for investment projects of industrial plants is formulated in the terms of decision theory. Presented approach is based on the reduction of multicriterion problem of investment project estimating to one-criterion problem. Application of simulation and artificial intelligence (AI) methods for estimation of technological and structural decisions, is the central feature of presented approach. Designed simulation model refers to discrete-event class. Object-oriented approach was applied for designing of the model and programming language C++ for its implementation. After it’s successful implementation this model was improved, in order to obtain robust, adaptive, re-configurable and responsive system. That was done with self-organization, self-learning, self-adaptation, self-optimization. This approach is based on application of AI methods (multi-agent systems (MAS), in particular) for simulation model development. MAS is a bio-inspired paradigm that allows to design systems based on autonomous and cooperative agents, exhibiting modularity, robustness, flexibility, adaptability and re-configurability. Presented methodology is tried-and-true method which applies on the phase of designing and engineering of industrial plant. This approach is discussed on the example of foundry plant with moulding line.