Optimization of Production in a Brewery using Artificial Neural Network

  1. Home
  2. Articles

Optimization of Production in a Brewery using Artificial Neural Network

Authors: Akaninwor, Godson Chijioke, Isaac, Okwu E. and Nkoi, Barinyima

Abstract

The aim of this research was to optimize production in a brewery using Artificial Neural Network (ANN). The Ishikawa diagram was also employed to ascertain the cause of low profit and its effect on the company. Low profit in the brewery was due to production of insufficient quantity of bottled bear, high cost of operation, more resources allocated for other products, frequent damage of products during the packing stage. Here, Feed-forward ANN was trained using set values of solved linear programming problems. In carrying out training for the network, the objective function was set with two (2) variables for canned and bottled beer and four (4) constraints equations for Malted Barley, Sorghum, Hops and Yeast. In the end of the training, the trained neural network was adopted for optimization of production profit for the 2 categories of beer. When the process was optimized, one (1) and fourteen (14) variables of the neutral structure network was formed, an output layer and 1 hidden layer produced 50 neutrons. Also, the feed-forward back propagation algorithm was adopted with a Bayesian regularization error. After producing 115,000 bottles and 137,000 cans of beer for the month of January at a volume 236.55 hectolitres (hl) and production cost of N24,673,333, gross profit was observed to be N41,653,333 with net profit of N16,980,000. In the end of ANN procedure, maximum optimization profit rose from N16,980,000 to N21,958,263.43 which makes up 29.32% increase. The regression plot for comparison of Simplex method profit and ANN profit gave a reasonable correlation of 0.92735 and goodness of fit/correlation coefficient up to 0.85998. Hence, the neural network correlation results proved to be 92% accurate, when compared with the Simplex method of solving linear programming problems. Thus, this research, if implemented, will lead to an increase in the production rate of beer being recorded, which will eventually bring about an increase in production profit on a monthly basis in International Breweries Plc, Port Harcourt.