Shixin Liu

An instance generator for project scheduling problems with multi-skilled personnel constraints

As the knowledge-based economy society comes, knowledge workers, which has replaced capital, facilities, raw material and other kinds of traditional resource, is becoming the most important resource in economic activity. Project scheduling problems with multi-skilled personnel constraints (PSPMSPC) has come to one of the research hotspots in project management field. Researchers have proposed a wide variety of models and algorithms, but there are not benchmark instances for testing these models and algorithms, which leads to the hardship in the performance comparison of different models and algorithms. Making use of the typical resource-constrained project scheduling problems (RCPSP) as basic instances, combining with the feature of multi-skilled personnel resource, this paper presents several parameters concern about the solving difficulty, and proposes an instance generator for the PSPMSPC. The proposed generator can generate instances with high randomicity under different problem scales and parameter sets, which has a high level of theoretical and practical value.

Research on dynamic pricing model of electronic airplane ticket in electronic commerce

The distinctive characteristics of electronic airplane ticket make its pricing method different from traditional pricing method. According to the different demands of passengers about the seats, electronic airplane ticket was sorted into multi-class. The income of each class of electronic airplane ticket was analyzed, considering its click number online and probability of passengers purchase. For the sake of maximizing the gross revenue of electronic airplane ticket, real-time multi-class dynamic pricing model was built at revenue management point of view. Then the state parameters in the math model were analyzed according to the operation status of Airline Company in practice. The recursion of the math model was deduced and proved. The results of experimentation verified that the methods are practical in computational efficiency and more close to practice.

Urban emergency intelligent decision system based on variable precision graded rough set on two universes

In recent years, different kinds of rough set models on two universes have been given. In this paper a variable precision graded rough set model based on two distinct but related universes (VPGRSTU) is proposed. We determine the scope of the grade from the absolute quantitative point of view, and the precision of n can be changed with user requirements. We introduce an algorithm to compute the new lower and upper approximations in VPGRSTU. Furthermore, a method to measure roughness of VPGRSTU is given. The method also can be used to compute the lower and upper approximations. An algorithm based on matrix is proposed to realize above method. Finally, several examples are employed to demonstrate the VPGRSTU and the theory in this paper. At last we build urban emergency intelligent decision system based on VPGRSTU.

Intelligent collaborative urban traffic management system based on SOA and cloud computing

Now the congestion of road has impacted on the city development seriously, and became the crux which constrained the city development. An intelligent urban traffic management system is needed urgently. In this paper an architecture of the intelligent collaborative urban traffic management system based on SOA and cloud computing is proposed. A collaborative service model based on SOA is given to realize intelligent collaborative management of urban traffic. The technology of service request and response dispatching based on SOA is proposed in the paper to match the collaborative event processing. Mass calculation was realized by the application of the cloud computing platform. The system fundamentally realizes the intelligent monitoring and management of urban traffic and realizes the purpose of intelligent dredge of urban traffic.

Multi-agent based simulation system modelling for steelmaking-continuous casting production process

The production process of steelmaking-continuous casting is a complex process which is high temperature, continuous, and uncertain. A multi-agent based simulation model is proposed to simulate the complex production process. Firstly, we present a simulation system architecture based on multi-agent technique. Secondly, we present the functional structure of agent and the communication / storage mechanism between the agents under this system architecture. Thirdly, we establish a hybrid model of multi-agent / discrete event dynamic system for simulating the steelmaking-continuous casting production process. The results of simulation experiments using practical production data of steel plants show that the model is valid and efficient.

A constraint programming-based branch and bound algorithm for job shop problems

In recent years, constraint programming (CP) has been widely applied to solved scheduling problems. As one of the key elements in CP, constraint propagation has been proved to be an efficient methodology to speed up the search process. In this paper, we develop a hybrid branch and bound algorithm with CP technique, called HCPBAB, which focuses on branches cutting to achieve outstanding performance tuning and constraints handling ability compared with the original method. Moreover, we extend the popular input/output negation, input-or-output constraint propagation process into HCPBAB to extensively test our proposed method. Experimental results on 40 instances taken from standard job shop benchmarks show that 33 instances are optimally solved while HCPBAB obtains 11 better solutions compared with Rego Cs results in 2009.

NSGA-II for slab selecting and reheating furnace scheduling in hot rolling production

Slab selecting and reheating furnace scheduling play a key role in reducing energy consumption and production cost. According to the production processes and constraints in the hot rolling production process, a kind of multi-objective optimization problems is presented to model the slab selecting and reheating furnace scheduling problems. A NSGA-II algorithm is proposed to solve this kind of problems. By being embedded a local searching procedure, the performance of the proposed NSGA-II is obviously improved. Computational experiment results of practical instances show that the proposed NSGA-II is effective in solving this kind of problems.

Hybrid model of multi-agent and DEDS for steelmaking-continuous casting-hot rolling manufacturing process simulation

The planning and scheduling problems of steelmaking - continuous casting - hot rolling (SCH) manufacturing process are hard to solve with traditional optimization approaches. As for the limitation of human ability, manual plans and schedules can not guarantee quality, robustness and material balance between processes. With the rapid development of computer simulation theories and virtual reality technologies, computer simulation has been becoming an effective method for solving complex problems and evaluating complex systems. Computer simulation system of SCH manufacturing process can be used for evaluating quality and robustness of plans and schedules in order to find the weakness and bottleneck in the manufacturing process before they are released. In this paper, a hybrid simulation model of multi-agent and discrete event dynamic system (DEDS) for SCH manufacturing process is proposed. The general simulation procedure and computer simulation system framework based on the hybrid model are presented. The structure, attributes, functions, interfaces, communication protocols and negotiation mechanism of different types of agent are defined. Based on the hybrid simulation model, a computer simulation system for SCH manufacturing process is developed in C++ language. The results of simulation experiments using practical production data of a Chinese steel plant show that the simulation model and the simulation system are valid and efficient.

Urban emergency intelligent decision system in rough set on two universes

In recent years, much attention has been given to the urban emergency intelligent decision system. In this paper, an urban emergency intelligent decision system in rough set on two universes is proposed. We give the structure of the urban emergency intelligent decision component. The binary relation R is redefined in this paper. We also defined the emergency system and emergency decision system. A model of intelligent decision in rough set on two universes is proposed. A realization method of intelligent decision based on matrix is given, and an algorithm is proposed to build the matrix. At last an illustrative example is employed to demonstrate the concepts, and methods.

A scatter search for hybrid flow shop problems with transportation

This paper studies a kind of hybrid flow shop problem with transportation delays (HFSTD) between consecutive production stages. A scatter search (SS) algorithm is proposed to solve the scheduling problem. A two-point order combination operator is designed for the SS. The authors randomly generate 12 instances to test the performance of the proposed algorithm. Experimental results show that the SS algorithm is efficient and effective for solving the HFSTD scheduling problem.