The efficiency and quality of current collaborative discussion for problem-solving can hardly satisfy the practical demands. To solve this problem, a formal model of collaborative discussion for problem-solving is established. First, to express thinking transformation in the process of collaborative discussion, the model of PTMM (Participant-driven two-directional mind map) is proposed, and then the formal definitions and operations for PTMM are provided. Second, the solution-space theory of PTMM is investigated. To let the solution be clearer, the concept and generation process of STMM (Solution-driven two-directional mind map) are given, which are derived from the PTMM. Third, inspired by the rhombus thought in extension theory, the RSM (Rhombus sequence model) based on PTMM and STMM is obtained and shown dynamically by visual means. Lastly, a practical example of collaborative discussion is shown, demonstrating that the RSM is able to effectively enlighten the thinking and promote the efficiency together with quality of collaborative discussion.

With the emergence of plenty of online cooperative discussion platforms,people can participate in the discussion without the limit of districts,and the content and profoundness of the discussion has increased dramatically.It is important to quickly search the core and hot topic from one discussion with lots of redundant information so as to accelerate the process of the discussion.This paper puts forward a method of analyzing the influence degree among all points of view in the whole process of discussion.By using the method of semantic similarity analysis based on Hownet semantic dictionary,all viewpoints in one discussion are automatically classified.Then the real-time influence degree of all these viewpoints is exhibited by using the visualization method.By this means,people can control the core and hot topic in the discussion.

Cooperative work logs can assist workers in disperse space to share cooperative work information. But there is just work flow control for work logs in currently cooperative work model. At the same time, the show of cooperative work logs is limited by chronological order. In this case, the staff of the cooperative project cannot recognize the relationships of numerous cooperative work logs and the controller cannot fast positioning current core work step. For mining the semantic relationships in cooperative work logs, this paper promotes a novel method, a visualization method of dimension reduction of log text semantic dimensional feature space, which help co-workers arranged in different parts of the cooperative work see the whole semantic structure clearly and assist the controller fast positioning the current core step. Thus, the controller is able to make a strategic decision for further work steps. In the end, the prototype of visualization platform is applied in the practical cooperative design of wire harness and proves this method was effective in fast positioning key emphasis in cooperative work by visualization of semantic relationships in cooperative work logs.

Focusing on the merit of process parallelism existing in the actual production,a job-shop scheduling model for parallel processes is proposed in this paper which aims to minimize makespan and considers machine flexibility among production processes.This model is implemented with a genetic algorithm-based scheduling algorithm in which a blocking encode method is used and a corresponding decode measure is put forward,which is suitable for parallel processes.The experiment result shows that the scheduling algorithm can solve the job-shop scheduling problem with parallel processes effectively.

With the intensive study of nonlinear system,people have learned that complex nonlinear mechanism existed in the interior of most phenomenon of world.Only relying on the abstract thought to understand law of things makes us feel more difficulty than ever before.In order to show the information laws hidden in the data stream of nonlinear system,based on the phase space reconstruction and manifold learning algorithm,a visualization method for analyzing the inherent law of nonlinear system was proposed;On this basis,the algorithm was improved which could only deal with the static time series data,a visualization method based on data stream was put forward,and the development trend of the law of nonlinear systems was dynamically displayed.After simulation,the results show that the approach can support people＇s understanding efficiently and analysis of nonlinear system.

Mind Map is an effective graphic tool in expressing divergent thinking and has been widely applied in brainstorming, work plan, decision making and document drafting and so on. The modeling process of Mind Map is based on humans' subjectivity and lack of logical criterion support. On the other hand, it is expert in divergent thinking but weak in convergence, so it cannot meet the needs of divergence-convergence in cooperative discussional problem solving. Accordingly, the concept of Extended Mind Map (EMM) is put forward and the description of its semantic relation and logical operations are given. Besides, some relevant visualization methods which are applicable to cooperative discussional problem solving are also introduced into EMM, so that it can classify concepts quickly, enlighten thinking and provide support to evaluate and converge the results contemporarily. In the end, the prototype system is applied in practical problem solving of wiring harness enterprises and proved to be effective.