Research On Encoding And Decoding Methods For Computing With Words Based On Type ? Fuzzy Sets

In view of present situation of mass data and tiny information in oilfield production,it has great theoretical significance or practical application value to realize intelligent management of oilfield water flooding production and improve recovery ratio by means of computing with words(CWW)method to construct intelligent evaluation system of oilfield production.In the paper,on the basis of the theory of interval type II fuzzy sets(IT2-FSs),and the encoding and decoding methods are studied of intelligent evaluation system of oil field injection-production well-pattern based on CWW.At first,the encoding methods are studied for the related descriptive words with geological information in oilfield production.The enhanced interval approach is proposed based on the theory of IT2-FSs,which is composed of data processing and fuzzification.In data preprocessing section,the processing conditions are added for interval endpoints,which could filter more reasonable intervals.Meanwhile,during outlier processing and tolerance limit processing,the interval endpoints and interval length are dealt separately,which reduces the amount of computation.The more exact intervals are remained at last.In fuzzification section,the method is enhanced for computing the lower membership function of interior footprint of uncertainties(FOUs)is modified,so the obtained FOUs models can distinguish more distinct words having close meanings.Finally,the words and corresponding FOUs models are compiled into a codebook,which is the basis of CWW engine and decoding.In order to obtain the optimal output of CWW,the similarity measure is adopted to decode the output of CWW engine.Based on the theory of IT2-FS,five similarity measures are investigated to determine whether they are satisfied with the four nature of reflexivity,symmetry,transitivity and overlap simultaneously.Theoretical verifycation and simulation show that the Jaccard similarity measure based on IT2-FS is suitable for the decoder because of its reliable output and fast computation.At last,combined with oil field application,the related descriptive words in the injection-production well pattern are encoded using the enhanced encoding method,and the codebook is generated.Based on useful rule bases and the output of perceptual reasoning obtained from earlier research,the outputs of CWW Engine are decoded by the means of five similarity measures so that the FOUs models are returned to words and the reasonable recommendation is given.Finally,the results of CWW are proven to be reliable and effective combined with the experts' experiences and the intelligent evaluation system of oil field injection-production well-pattern are constructed completely.