Research Simulation Of The TRIZ Contradiction Matrix Based On SVM

TRIZ is short for the Latin transliteration for the Russian phrase of "Theory of theSolution of Inventive Problems". It was a theory started by the former Soviet inventor andscientist, Genrich S. Altshuler, for the solutions to invention and innovation problems. Withthe collapse of the Soviet Union in1990s, the theory was opened up to the world and attractedgeneral attentions and research. Nowadays it has been adopted by a number of world-famedmanufacturing enterprises, and has contributed them much on their production quanlitypromoting and technology advancing.The Contradiction Matrix is the principal method of the TRIZ to solve technicalcontradiction problems in the Industrial Engineering. The method is proposed by the founderof the TRIZ after a deep research and summary on over2.5million invention patents,covering39engineering parameters involved in the common technical contradictions and40invention principles aiming at the parameters. The method can be regarded as the guidingdirections and evaluation basis for the innovation, by which can rapidly and effectively reachthe optimization of the innovation.The traditional method to exploit the Contradiction Matrix is manual look-up. In reality,the technical contradiction problems always appear accompanied by multiple pairs ofcontradictions, thus the hugeness of the matrix and the complexity of the problem definitionhas increased the workload and the difficulty of the manual look-up.As the mapping relations among the elements in the Contradiction Matrix being explicit,it is appropriate to proceed a training on the matrix with the machine learning methods. Themapping relations among the elements are non-linear mappings, in addition the amounts ofthe elements are appropriate moderate for the machine learning process, therefore this paperadopts the Support Vector Machines (SVM) as the machine learning method and conducts asimulation of the Contradiction Matrix, thereby achieves the automated solutions of thetechnical contradiction problems instead of manual look-up with the Contradiction Matrix.After a programming experiment, the simulation turned to be successful. To input thecontradiction parameters of the matrix can obtain the invention principles in accord with theparameters as the output. The successful implementation of the simulation of the TRIZContradiction Matrix based on the SVM facilitates the progress of the TRIZ’s practicability,provides a reference for the developing of the computer-aided innovation designing softwaresbased on the TRIZ, contributes to the popularization of the application of the TRIZ, whichcan help to promote people’s comprehension on the innovation and their innovation abilities.