Study Of Bidding Decision-making In Construction Project

With the regularizing of the construction market in our country and the enlarging of the exchange between China and foreign countries after our country joins WTO, our construction firms face unprecedented competition pressure. To win an ideal project for the construction firm in a fierce market dependents both on the good business capability of the firm and its agency, and on the accurate calculation on the bidding, especially the ascertain of the mark-up, which plays an important part in that whether the contractor can win the bidding on reasonable price, and the existence and development of the construction enterprise in the future. To help our construction firms to make the best bidding decision according to the different project situations, the dissertation will improve and develop the conventional bidding decision theory of construction project under the lowest price bidding. The main work and innovations of the dissertation include:1. To overcome the disadvantages of slow convergence speed a, being prone to converge to minimum and hidden layer neuron number hard to be determined, Genetic Algorithm (GA) was used to combine with the BP neural network. A bidding model for the construction project based on the BP neural network improved by GA was put forward. It has been proved by an application as an effective complement for the deficiency of BP neural network on the bidding decision of construction project.2. When an input data set is large, the adaptive-network-based fuzzy inference systems (ANFIS) has the disadvantages of slow learning speed, no convergence, and being disturbed by redundancy. So in this dissertation rough set theory (RS) was applied to preprocess the input data set. Then a new Bidding Decision-Making model based on RS and ANFIS was established for the determination of mark-up of the construction project. It is concluded that the model by this method is smaller in network scale, clearer in network structure, and faster in learning speed than the normal BP neural network. Moreover, this method remains high accuracy.3. According to the dispute about that whether it is stochastic independence for the contractor to win over a number of competitors between the two traditional bidding models of construction project: Friedman's model and Gates'model, this paper presented a novel bidding method: Monte Carlo method. The case calculation results show that the Monte Carlo method can avoid the dispute between Friedman's model and Gates'model and get more profit than Friedman's model and Gates'model.4. In imperfect massage repeat game, bidder's quote will fall into"the prisoner's dilemma". Then the recent study on bidding under the condition of imperfect massage mainly focuses on the single stage game. So in this dissertation, the imperfect massage repeat game: KMRW reputation model was applied to study the tenderer's bidding decision. The research results show that if the probability of every tenderer being irrational is positive, no matter how small the probability is, the cooperative will be equilibrium in finite repeat games between two tenderers if the number of repeat games is large enough.5. In the present research of bidding decision for construction project, the tenderee's preference on construction products is usually neglected. So, this dissertation puts forward that the construction product differences should be put into the bidding decision for construction project. Then the Nash equilibrium of bidding decision between two tenderers was studied based on the Hotelling model under the condition of construction product heterogeneity. The research results show that the larger the extend of the quality differentiation of the tenderer's construction products is and the higher the tenderee's preference on the quality of the construction products is, the larger expected equilibrium payoff the tenderer can get.