Studies Of Hybrid Multiple Attribute Group Decision-making On The Backbone-network Reconfiguration Schemes

Modern society is increasingly dependent on reliable power supply. In order to optimize the allocation of resources and energy, power systems are continuously interconnected and expanded. Dynamic behaviors of power systems are becoming more and more complex. Blackouts may be caused by improper treatment of local outage. The restoration of power system is attracting more and more attention for the catastrophic results of blackouts. Blackout is inherently inevitable because of various reasons beyond prediction, unforeseen circumstances and its self-organized criticality.Once blackout occurs, it is imperative to take effective and secure measures to restore the power system as soon as possible. In backbone-network reconfiguration, many conflicting factors should be considered comprehensively to select a compromise solution. As the core function of decision support system, the optimization of decision-making is indispensable. Optimization of restoration decision-making is meaningful for guaranteeing fast and safe restoration, providing assistance to decision makers as well as improving restoration strategies. Based on literature review of power system restoration and the abovementioned problems, optimization of decision-making in power system restoration is studied. The main research results are as follows:(1) The principles and procedures of establishing the evaluation index system is described, the evaluation index system is established based on the process and characteristics of backbone-network reconfiguration, the calculation of attributes values and type of attributes are analyzed in detail. For backbone-network reconfiguration, the indices related to system security, restoration speed and restoration benefits can be used as attributes.(2) A sequential screening method is proposed to simplify the decision-making process. The number of alternatives is reduced sequentially by simple linguistic screening techniques, Pareto optimality based screening, and Pugh matrix. Thresholds are set for each attribute to screen out extreme restoration schemes. Dominated solutions are deleted to retain Pareto-optimal solutions. Attributes and restoration schemes are compared mutually for rough qualitative analysis, the procedure is simple but structured, weak solutions are determined rapidly, the influence of the extreme restoration schemes are reduced, the computation burden and simulation time are decreased.(3) A hybrid multiple attribute group decision-making method based on VIKOR (in Serbian:Visekriterijumska Optimizacija I Kompromisno Resenje) is proposed to comprehensively evaluate and rank restoration schemes. Both qualitative and quantitative criteria are considered simultaneously, the uncertainties and subjectivity in decision-making can also be counted. It can reflect the vagueness and uncertainties of restoration problems, and avoid too much fuzzification. Different forms of preference relations in group decision-making are integrated to obtain consistent subjective preferences. Objective weights based on distance deviation are integrated with subjective weights by Quadratic programming model to obtain combined weights, the computation time is reduced greatly. VIKOR is chosen because it considers not only maximum group utility, which reflects best performance of all criteria, but also individual regret, which represents the most unsatisfactory attribute, the ratio of the two indexes can be adjusted flexibly. At last, sensitivity analysis on weights, parameters and attribute values is defined to guide the DMs, If time allows, the stability intervals of attributes can be obtained, and in an emergency, qualitative analysis can be conducted to determine the sensitivity.