According to the problem that there have not any research focuses on building a general crowdsourcing platform that can be serviced any enterprises. This paper have studied the key issues such as the platform framework, operation mode and construction principles and other issues to be considered when building such a special crowdsourcing platform. The crowdsourcing platform that build based on the key issued studied in this paper has greater practical value, which can be used as the communication bridge between crowdsourcers and contractors and service all the enterprises and the community, thus achieving a win-win goal between the crowdsourcers and contractors.

How to ensure a smooth, fast and efficient emergency response procedure becomes a highly concerned issue. However, a procedure of emergency plan may be confusing and inefficient in reality due to delay caused by waiting for decision-making, responding to conflicts and limited resource during the process of dealing with emergency. In this paper, we propose a colored stochastic Petri net to evaluate the security and complexity of emergency response procedure and the reasonableness of resource flow, so as to effectively analyze the potential deficiencies of emergency response procedure. We firstly establish a colored stochastic Petri net, and then convert the colored stochastic Petri net to an isomorphic Markov chain. Studying the structural properties of the colored stochastic Petri net and steady-state nature of the Markov chainprovides a scientific basis for the perfection of emergency plan. Meanwhile, it also ensures an ordered and efficient implementation of emergency response procedure in an emergency.

Due to the emergency plan format information characteristics is stronger and more easily to gain the complete content description, so this paper puts forward the recommendation algorithm of emergency plan based on content filtering. This paper proposes the characteristics model of emergency plan and the user's interest model and designs the flexible recommendation interface for user interactive, so we can get the recommend results.

The core of exam management system is how to generate test paper automatically. The existing algorithmic is difficult to simultaneously achieve the requirements with efficient, random, flexible and expansive performance. In this paper, we introduce the maximum flow algorithm of the upper and lower bounds in the graph theory, create a test paper generation model based on constraint conditions, and implement the test paper generation under complex constraints. In addition, to illustrate the success rate and efficiency, the system generated automatically test papers on proposition needs. In turn, it verifies the correctness and rationality of the model. The algorithm proposed in this paper will provide theoretical and technical support for other systems, and it further promotes the development of the exam management system.

Nowadays, automatic scoring is an important way of teaching and examinations. However, there is no existing research on automatic scoring for subjective item of database domain both at home and abroad. According to the characteristics of database domain, we construct database domain synonyms ontology and proposed a text similarity calculation algorithm based on Hamming distance. Then we implement the automatic scoring for subjective item of database domain on the basis of ontology. In addition, in order to verify the accuracy and rationality of the algorithm, we take a specific subject as an example. The experiment results further illustrate the accuracy and efficiency of the proposed automatic scoring algorithm.