| The environmental problems have drawn academia’s attention. Among them, environmental performance evaluation and quota allocation are two crucial issues. Data envelopment analysis (DEA) is a common used mathematical programming tool to evaluate the relative efficiencies of decision making units (DMUs). Also, it can be employed for ex ante plan. As the sums of undesirable outputs are always limited by decision makers, zero sum gains (ZSG) DEA model is a reasonable choice for environmental performance evaluation. In addition, the uniform frontier and centralized allocation models can be applied for quota allocation of undesirable outputs.However, the ZSG-DEA models are nonlinear, which brings up bottleneck in real-life application. In addition, the uniform frontier can only allocate single undesirable output and it pays no attention to the overall interest, while the centralized models concentrate only on overall interest and ignore individual interests. To remedy these defects, this thesis accordingly develops new models and methods to evaluate the environmental performances and allocate quotas of the undesirable outputs for different DMUs.This thesis involves six chapters, and the essence is summarized as follows.Chapter1introduces the motivation and the major contributions of this thesis.Chapter2firstly summarizes the fundamental theory of DEA, which includes the basic concepts, basic and extensive models and theoretical applications. Secondly, the relevant works are reviewed associated with DEA-based environmental performance evaluation and quotas allocation. Finally, based on the literature review, some unsolved problems are summarized.Chapter3firstly transforms the existing ZSG-DEA models into linear or parametric linear ones. Secondly, it is shown that the linear formulations of some ZSG-DEA models are equivalent to super-efficiency models. As a matter of course, the models may encounter infeasibility under VRS assumption. At last, the linearized models are adopted to evaluate the environmental efficiencies of30administrative regions in China.Chapter4is an extensive work of Chapter3. It proposes two improved ZSG-DEA models and an available method to obtain the efficiency values of some infeasible DMUs under VRS assumption. Based on the proposed models, an iterative algorithm is devised to generate the uniform frontier with multiple ZSG inputs (outputs). Finally, an empirical study of allocating the quotas of three industrial wastes for30administrative regions in China is presented.Chapter5still concentrates on the quotas allocation of undesirable outputs. The existing allocation plan is always centralized and will suffer from an implementation difficulty. While the uniform frontier in Chapter4pays no attention to overall interest. In this chapter, a two-step method including the centralized allocation model and compensation scheme is provided to allocate the carbon emissions abatement (CEA) among different DMUs, which can consult the overall and individual interests in the meantime. In the empirical example, the proposed method is applied to the CEAs allocation for the countries in Organization for Economic Co-operation and Development (OECD).Chapter6draws the remarks and conclusions of this thesis and provides some available directions for future researches. |