Policy and Complex Systems
In the progress of science , there is some knowledge that is not superseded or nullified , no matter how many publications are published . The core knowledge is no longer cited . Thus , the box “ knowledge ” in Figure 4 assumes that some core human knowledge remains stable . The model shows causal relationships among variables whose interaction creates system dynamics . The system determines a behavior regarding scientific knowledge growth . The relationships between knowledge growth and its behavior can be explained in two structural features — stock and flow , and feedback loop . First , the knowledge growth can be explained as a function between net creation rate and initial knowledge as seen in Equations 1 and 2 . The input of new scientific knowledge creates growth of scientific knowledge in the system which is stored , and it replaces other scientific knowledge in the Popperian sense of eliminating error or confirming earlier findings . Thus , some knowledge experiences ‘ obsolescence ’ in the sense that it is no longer cited even though a small percentage becomes part of normalized knowledge . The growth of knowledge can be assumed to influence the obsolescence rate ( this can also be a proxy for lower quality work ) with crowding increasing the rate . In this regard , the growth of knowledge can be determined by net creation rate and initial stock of knowledge . The net creation rate is determined by the gap between creation and obsolescence rates . Knowledge = INTEGRAL ( net creation rate , initial knowledge )
The net creation rate is determined by the gap between creation and death rates .
net creation rate = creation rate ( cK ) - obsolescence rate ( oK ) = ( c - o ) a * K where c is a creation rate , o is an obsolescence rate , and a is a power value of the relationship between c and d , and K is amount of knowledge . Specifically , the pattern of scientific knowledge growth , pure exponential growth , or exponential growth with saturation can be determined by the pattern of net creation rate as seen in Figure 5 . If creation dominates obsolescence ( c > o ) linearly ( a = 1 ) or nonlinearly ( a > 1 ), scientific knowledge will show pure exponential growth . However , if creation dominates death , and then death dominates creation like the bell-shape as seen in Figure 6 , the pattern of scientific knowledge growth will show an s-shape growth ( here , an exponential growth with saturation ).
Scientific knowledge growth can be identified along with various feedback structures or loops around stocks and flows . Feedback structure or loops means a closed causal circle among variables . In general , there are two types of feedback structures : positive ( self-reinforcement ) and negative ( balance ). Positive feedback structures are self-reinforcing processes wherein action creates a virtuous circle . Negative feedback structure , on the other hand , is a process to stabilize or balance a system . Thus , behaviors in a system become determined by a type of feedback loop in the system . When a positive feedback dominates the whole system , the system tends to show an exponential growth . When negative feedback dominates , the system tends to show an upward or downward convex growth . An s-shape growth in a system tends to be formed when positive feedback dominates , followed by when negative feedback dominates .
In our system dynamics model , there is one positive feedback structure and two negative feedback structures as seen in Figure 5 . The first positive feedback structure ( R1 ) is the feedback loop from knowledge through creation back to knowledge .
116