No abstract

No abstract

No abstract

Static model of the distillation column installation of primary oil refining is designed. The algorithm of BP-method in mathematical package Matlabis realized. For representing the composition of the oil, kind of curve fitting the true boiling point (CTI) of oil from the experimental data is determined.

Simulation of queuing systems (QS) is an effective tool for researching of production systems. Many foreign [1, 2] and Russian [3, 4] sources show the effectiveness of use of simulation in a study and optimization of activities of enterprises and their subdivisions. In particular, it is shown that queuing and multi-agent modeling is an effective in a decision support systems of a production management [5]. Let's notice that the multi-agent system (MAS) is a simulation model in which objects ("agents") are characterized by some autonomy, decentralization and limited ideas [6]. Extremely autonomy and limitations are characteristic of channels and transacts of QS. Software agents of MAS are described by algorithms of their own actions and interactions with other software agents. Statistical distribution of random variables has a great influence on the modeling results. As shown in [7] a model of joint works of titanium plant subdivisions gives very different results in case of uniform or Poisson distributions (Fig. 1). The specific units of reaction and factors do not matter in this case. The non-linear relationship between the reactions with normal and Poisson distributions of the factor is important. This dependence is also characteristic of other queuing systems. The uniform distribution is "worst" in some manner. It provides biggest deviations of a result value of any QS. Meanwhile the uniform distribution is assumed by default exactly when no reliable information in an input data. The Gaussian distribution could be used customarily when it can be confirmed by statistical testing of the hypothesis. Using of conclusions obtained solely through the use of simulation programs (QS) is considered insufficient in qualifying scientific jobs (thesis). Famous methodologist of dissertation research in Russia considers [8] that "…an idea and technique of simulation is so simple that you can learn it of some days in a matter. Therefore, use of them … cannot be a confirmation of the high qualification of a thesis". The scientific tradition requires supplementing QS calculations with analytical research.