The impact of network topological structures on systematic technology adoption and carbon emission reduction

This paper investigates how the topological structure of the technological spillover network among agents affects the adoption of a new clean technology and the reduction of system’s carbon emissions. Through building a systematic technology adoption model with technological spillover effect among agents from the network perspective, this paper first illustrates how the new technology diffuses from the earlier adopters to the later adopters under different network topological structures. Further, this paper examines how the carbon emission constraints imposed on pilot agents affect the carbon emissions of other agents and the entire system under different network topological structures. Simulation results of our study suggest that, (1) different topological structures of the technological spillover network have great influence on the adoption and diffusion of a new advanced technology; (2) imposing carbon emission constraints on pilot agents can reduce carbon emissions of other agents and thereby the entire system. However, the effectiveness of the carbon emission constraints is also largely determined by the network topological structures. Our study implies that the empirical research of the network topological structure among the participating entities is a pre-requisite to evaluate the real effectiveness of a carbon emission reduction policy from the system perspective.


Supplementary Methods
The systematic technology adoption model for each agent Each decision agent minimizes the total cost over its decision-making period with the following objective function (S1): where denotes the discount rate, and T denotes the length of one agent's decision-making period (i.e., its foresight).
The first part of the objective function represents the total investment cost of all three technologies, where, denotes the newly installed capacity for technology i at time t; denotes the unit investment cost for technology i at time t, which for T1 is a constant. For T2 and T3, since we assume that they have learning potential, their unit investment costs are calculated with the following Eq. (S2): where, 0 denotes the initial unit investment cost for technology i; 1 − 2 − denotes the learning rate of technology i, which indicates the cost reduction percentage when the cumulative production of technology i doubles; is the progress ratio; � represents the cumulative production of technology i at time t, which is calculated with the following Eq. (S3): where ̅ 0 denotes the initial cumulative production using technology i before the firstdecision period; denotes the production using technology i at time j.
The second part of the objective function represents the total resource extraction cost, where denotes unit resource extraction cost at time t, which is computed with the following Eq. (S4): where 0 denotes the initial unit resource extraction cost; is a resource extraction cost coefficient; ��� denotes the cumulative resource extraction by time t : where is the total resource consumption at time j. It is the sum of the resources consumed by all three technologies at time j : where denotes the efficiency of technology i.
The third part of the objective function is the total operation and maintenance cost, where denotes the unit OM cost for technology i.
Inequation set (S8) indicates that the production using one technology i cannot go beyond its total installed capacity at each time t. Here, denotes the total installed capacity of technology i at time t, which is computed with the following Eq. (S11): where, 0 denotes the initial total installed capacity of technology i before the first decision period; is the plant life of technology i.