Intelligent manufacturing strategy selection
The pdf file stated that “Dynamic Causal Mining (DCM) is applied in this work to extract knowledge to assist strategy selection in the form of causal rules among historical data of strategies” then:
there is a place for innovation in this strategy?, and in affirmative case how is applied.
Thank you for this question. The paper only reports our preliminary thoughts on this subject. Work is being conducted to compare the game-theory approach and other methods of selecting strategies. The results will be presented at an appropriate future event. Have a good time in Prague!
Thank you for your question. I suspect that, because we use past records to induce selection rules, the selection will follow well-trodden paths rather than being innovative. In our future work, we will look into the possibility of enhancing innovation.
This paper is very interesting for the combination of methods which are suggested, but also at a higher level since it sets a very important problem of artificial intelligence: trying to automate high level reasonings is always difficult, and most of the time, huge systems are built and are only capable of solving rather simple problems. Being a little bit provocative, in the case of strategic decision making, would not it be more realistic to work on how to feed the human decision maker with better input information than to try to automate his decision making, since the decision making is in my opinion the strong point of the human decision maker ?
The other strategy selection methods I stuydied is System dynamics. For the past decades Game theory has been dominate the enomical part of strategy selection, but recently system dynamics has been used in dynamic causal strategy selection.
And game theory has now been applied in chemistry and other areas.
System dynamics itself enhances innovation, although when added with data mining it limited the innovation, because the dynamic hypothesis is limited. However the causal rules extracted from the data can act as causal loops which show networks of cause and effect which can identify unintended consequences and underlying possibilities.
First of all I agree that it is difficult to automate high level reasoning, but that doesn't mean we should not try. The method suggested in this paper is used as a decision support and when there is no proper decision maker is could act as one and give an more appropriate decision.
And of course there should be a continues research to improve the system's input.However there are still a decision to make, weather it is made by human or machine.By having an independent automation of decision the decision maker could be more certain and more confident of the future decision.
I found your paper very interesting.
You mentioned the "Prisioner's Dilemma", an example of a cooperative game. But in manufacturing environments, the players have incomplete or asymmetric information.
What if even after using DCM the information is incomplete or asymmetric?
Due to the fact that the data gathered by diffrent company are different. The information will always be asymmetric. Even in real life there are asymmetric information everywhere. But with this methods the companies with no experience in the field might gain some extra information which will hopefully improve their performance.
Bonjour, Bernard! My apologies for the delay in responding to your comment. I had thought Yi had done it until I discovered that his reply did not quite address your comment. I agree with you that we should aim to assist the human decision maker in arriving at the best strategic decision rather than trying to automate the decision making process. The system proposed in our paper is indeed only a decision support tool. The final decision must still be taken by humans. Merci bien et bonnes vacances!
Have you been able to study and compare and contrast your method based on game-theory with otger ways of selecting a manufacturing stragegy?