Imacs 2016 imecs 2016 Proceedings (Preliminary version) of the 4


EXPERT SYSTEM FOR WAREHOUSE STOCK OPTIMIZATION



Download 2.73 Mb.
Page16/62
Date20.10.2016
Size2.73 Mb.
#5106
1   ...   12   13   14   15   16   17   18   19   ...   62

EXPERT SYSTEM FOR WAREHOUSE STOCK OPTIMIZATION

97.Radim Farana – Ivo Formánek – Bogdan Walek



Abstract

Currently, many companies try to optimize their system of warehouse stock management to minimize their production costs. The optimization means mainly optimization of processes like resources adjustment, resources planning, purchasing, deliveries, sales etc. The goal is clear – not to spend too much money for stock.

There are various information systems more or less successfully anticipating and predicting the quantity of resources that should be ordered. These systems usually apply prediction based on sales in the previous period. In most cases the managers can also adjust the prediction based on their knowledge and experience. But not every company has so experienced managers. Therefore it is advantageous to have something like expert system for this.

The paper will introduce an expert system which will use a knowledge-base based on managers’ knowledge and experience and other influences affecting the prediction in order to predict and propose the quantity of necessary resources that should be ordered. Individual steps and parts of the expert system are described in the paper. The proposed expert system is verified on a particular example, i.e. in practical application.


Key words: expert system, fuzzy logic, optimization, warehouse stock
JEL Code: C49, C53, C61

Introduction

Currently, many companies try to optimize their system of warehouse stock management to minimize their production costs. The optimization means mainly optimization of processes like resources adjustment, resources planning, purchasing, deliveries, sales etc. The goal is clear – not to spend too much money for stock. There are various information systems more or less successfully anticipating and predicting the quantity of resources that should be ordered.

There are generally used different approaches to the sales prediction and thereby the production planning (Brown, 2000; Swift, 2001). These can be equalized on statistical methods, especially the analysis of time series, but in practise we often come across with very simple approaches that are very robust at the same time, such as method of moving average. Our approach is based on the use of fuzzy logic expert systems (Novak, 1995; Pokorny, 1996). Experts systems, in particular using fuzzy logic are in this area used by a number of authors for different applications (Xu Bin et al., 2010; Zang et al., 2004; Zang, Liu, 2005). The applications of artificial neural networks (Vaisla et al., 2010; Kunwar et al., 2010) or tools of soft computing are also very interesting. As advantage of Rule-Based Expert Systems is a particular opportunity to use the knowledge of experts and their simple expressions by rules. Fuzzy logic then helps us especially with easy expression of dependences among the values which is poorly expressed using crisp values.


98.1 Design of the expert system for prediction of sales


As an example of the use of the expert system for prediction of sales we use the specific example of the sale of the three specific products of a mechanical engineering company. We have sales data for each of the weeks in 2014 and 2015 (Fig. 1), which are split into two parts. We have used the data 2014 and the first 42 weeks in 2015 to determine the knowledge. The remaining data, then we have used to verify the behaviour of the expert system compared to the prediction based on moving average.

Fig. 1: The progress of the sales of the three products of the engineering firm



Source: authors

The first task is to determine the parameters of expert system, because we have no other knowledge than the actual sale, we can't use a more comprehensive view of the problem, as in the other cases, see for example Walek, Farana (2015). Thus, as the parameters we set the sales of individual products in the previous two weeks. For the realization of the expert system, we will use the Linguistic Fuzzy Logic Controller (LFLC), see Fig. 3, which is very convenient for practical applications.

Linguistic Fuzzy Logic Controller (LFLC) is a result of application of formal theory of the fuzzy logic in the broad sense (FLb). The fundamental concepts of FLb are evaluative linguistic expressions and linguistic description. Evaluative (linguistic) expressions are natural language expressions such as small, medium, big, about twenty-five, roughly one hundred, very short, more or less deep, not very tall, roughly warm or medium hot, roughly strong, roughly medium important, and many others. They form a small, but very important, constituent of natural language since we use them in common sense speech to be able to evaluate phenomena around. Evaluative expressions have an important role in our life because they help us determine our decisions; help us in learning and understanding, and in many other activities.

Simple evaluative linguistic expressions (possibly with signs) have a general form (where is one of the adjectives (also called gradable) “small – sm, medium – me, big – bi” or “zero – ze”. The is an intensifying adverb such as “extremely – ex, significantly – si, very – ve, rather – ra, more or less – ml, roughly – ro, quite roughly – qr, very roughly – vr”), see Fig. 2.

This set of linguistic expressions has been drawn up on the basis of the experience of the creators, but it does not always suit the particular situation. Fig. 3 shows the frequency of each value of sales for the product P1. We can see that most of the values are concentrated in the middle of the interval, which covers little linguistic expressions, so when compiling a system of rules for the expert system, there have often appeared the same value (me). LFLC tool offers the possibility of user-set assembly of evaluative linguistic expressions; see Fig. 3 that will better respond to the current situation.

Fig. 2: A standard set of evaluative linguistic expressions of LFLC tool

d:\users\far10\texty\prispevky\2016\05_imecs\lflc-valuesstandard.bmp

Source: authors

Fig. 3: The frequency of sales of the product P1



Source: authors

Then we set the rules that describe the behaviour of the system, the sale of products. As already mentioned, in the sample we have no information about external influences, the behaviour of competitors, etc., that would have allowed us to better describe the whole system, which could significantly improve the quality of prediction, see e.g. Walek, Farana (2015), and we need to focus only on known values of sales.



Fig. 4: A set of evaluative linguistic expressions drawn up on the basis of expert knowledge

d:\users\far10\texty\prispevky\2016\05_imecs\lflc-valuesexpert.bmp

Source: authors

Fig. 5: A set of evaluative linguistic expressions drawn up on the basis of expert knowledge



Source: authors

We will use the two previous sales values (marked P1-2, and P1-1, P2-2 and P2-1, P3-2 and P3-1) and set the individual rules predicting the current sale (P1, P2, P3). If we do not have available the knowledge of an expert, we can use the automated system with advantage of learning system LFLC, which on the basis of known values shall draw up a set of rules, see Fig. 5.

The created set of rules, of course, can contain not only duplicate rules, but also contradictory rules, which it is appropriate to resolve, but the expert knowledge is required. If the contradictory rules are not removed, the system will determine the outcome on the basis of all these rules laid down by the way, typically by the method of Center of Gravity. Fig. 6 shows the testing the behaviour of the created fuzzy expert system for the specified combination of input values.

Fig. 6: Testing the fuzzy set of rules in LFLC environment

d:\users\far10\texty\prispevky\2016\05_imecs\lflc-test.bmp

Source: authors

As already stated, the set of rules has been drawn up on the basis of sales data from 2014 and the first 42 weeks of the year 2015. If we used the expert system for prediction, the results were satisfactory, but the deviation would be similar as the other approaches. Of course, the process of sales, Fig. 1 clears that the sale of the product P1 has a positive trend over the time. This fact we can very easily use for a simple change in the context of the input and output variables, see Fig. 7.



Fig. 7: Setting the context of the output variables P1 in LFLC environment



Source: authors

Tab. 1 shows the results of the testing for the solution for weeks 43-53 of year 2015. We can see that compared to the moving average (of the four previous values), the expert system reaches an average half a deviation from the true value.



Tab. 1: Title of table

Source: authors


99.Conclusion


The paper has presented a processing of design for the fuzzy-expert system for prediction of sales of products of the engineering company, using the tool of Linguistic Fuzzy-Logic Controller, which has been developed at the University of Ostrava in the city of Ostrava. The described procedure for creating a set of linguistic expressions and automatic assembly system of the rules enables to easily automate the whole process. A simple change of input and output contexts then allows to take into account the global development trend of sales and thus significantly advanced prediction using an existing set of rules.

Experience of using the expert systems prediction of warehouse stock shows that these systems work very well in situations with no unexpected changes or events. The main reason is that the set of rules contains only aggregated knowledge of situations that already occurred.. Therefore, inference mechanism of expert systems can find solutions just only for situations which are the same as situations occurred in the past or which are very similar to them. The presented example comes from middle-sized engineering company that did not monitor and evaluate the external influences and their effects on the sale. To further refine the prediction and to obtain more accurate prediction, there would have to contribute an analysis of external influences and add the appropriate parameters in the expert system, see Walek, Farana (2015).

Operation of the predictive system in early 2016 has confirmed our expectations with which it has been created. Any deviations in the predicted values continue to move in units of percent. These good results are achieved also by inclusion of new rules to the values of sales which is very easy in case of the expert systems – e.g. unlike artificial neural networks that need a new learning process.

100.Acknowledgment


This work was supported by the project “LQ1602 IT4Innovations excellence in science”.

101.References


Brown, Stanley A. Customer Relationship Management: A Strategic Imperative in the World of E-Business. New York: John Wiley, 2000.

Swift, Ronald S. Accelerating Customer Relationships: Using CRM and Relationship Technologies. Upper Saddle River: Prentice Hall, 2001.

Novak, Vilem. “Linguistically Oriented Fuzzy Logic Control and Its Design.” Int. Journal of Approximate Reasoning vol. 12, 1995, pp. 263-277.

Pokorny, Miroslav. Artificial Intelligence in modelling and control (in Czech). BEN - technická literatura: Praha, 1996, 189 pp. ISBN: 80-901984-4-9.

Xu Bin, Liu Zhi-Tao, Nan Feng-Qiang, Liao Xin. “Research on energy characteristic prediction expert system for gun propellant.” IEEE International Conference on Intelligent Computing and Intelligent Systems (ICIS), 2010, Volume 2, pp. 732 – 736, ISBN: 978-1-4244-6582-8.

Zhang Bofeng, Wang Na, Wu Gengfeng, Li Sheng. “Research on a personalized expert system explanation method based on fuzzy user model.” Fifth World Congress on Intelligent Control and Automation, 2004. WCICA 2004, Volume 5, pp 3996 – 4000, ISBN 0-7803-8273-0.

Zhang Bofeng, Liu Yue “Customized explanation in expert system for earthquake prediction.” 17th IEEE International Conference on Tools with Artificial Intelligence ICTAI 05, 2005, 5 pp. – 371, ISBN: 0-7695-2488-5

Wang John. Data warehousing and mining: concepts, methodologies, tools, and applications. Information Science Reference: Hershey, PA, c2008, 6 v. (lxxi, 3699, 20 p.). ISBN 978-1-59904-951-9.

Khosrow-Pour, Mehdi. Encyclopedia of information science and technology. 3rd ed. IGI Global, 2014, 10384 p, ISBN 978-1-46665-889-9 .

K S Vaisla, Ashutosh Kumar Bhatt, Shishir Kumar. “Stock Market Forecasting using Artificial Neural Network and Statistical Technique: A Comparison Report.”. (IJCNS) International Journal of Computer and Network Security, Vol. 2, No. 8, August 2010.

Kunwar Singh Vaisla, Ashutosh Kumar Bhatt. “An Analysis of the Performance of Artificial Neural Network Technique for Stock Market Forecasting.”. (IJCSE) International Journal on Computer Science and Engineering Vol. 02, No. 06, 2010, 2104-2109.

Walek, Bogdan, Farana, Radim. “Proposal of an Expert System for Predicting Warehouse Stock”. 4th Computer Science On-line Conference 2015, CSOC 2015. Zlín: UTB ve Zlíně, 27. – 30. 4. 2015, pp. 85-91. ISSN 2194-5357.

Contact

Radim Farana

Institute for Research and Applications of Fuzzy Modeling, University of Ostrava

30. dubna 22, 701 03 Ostrava, Czech Republic

radim.farana@osu.cz
Ivo Formánek

Department of Entrepreneurship and Management, University of Entrepreneurship and Law

Michálkovická 1810/181, 710 00 Ostrava, Czech Republic

ivo.formanek@vspp.cz


Bogdan Walek

Institute for Research and Applications of Fuzzy Modeling, University of Ostrava

30. dubna 22, 701 03 Ostrava, Czech Republic

bogdan.walek@osu.cz




Download 2.73 Mb.

Share with your friends:
1   ...   12   13   14   15   16   17   18   19   ...   62




The database is protected by copyright ©ininet.org 2024
send message

    Main page