TY - JOUR
T1 - ON SYSTEM COMPLEXITY, STABILITY AND PERFORMANCE
T2 - APPLICATION TO PREDICTION
AU - Ratsaby, Joel
N1 - Publisher Copyright:
© 2024 MSP (Mathematical Sciences Publishers)
PY - 2024
Y1 - 2024
N2 - Intuitively, the three properties of complexity, stability, and system performance are interrelated. For instance, more complex ecological systems tend to be less stable, and in software engineering, software performance reliability is directly affected by software complexity. We introduce a new rigorous framework that shows how these three properties interrelate. We focus on a specific family of systems that predict input binary Markov chains. A system’s output is a binary sequence that indicates when it fails to predict the input. System complexity is the average number of information bits needed to describe the output per input bit. System stability is the discrepancy between the system’s output when presented with two random input sequences. A bound on the prediction error is derived and used as a system’s performance guarantee. It is shown that as complexity increases, stability decreases, and performance guarantee becomes more sensitive to changes in the input, making the system less robust. The analysis taken here is applicable to more general systems.
AB - Intuitively, the three properties of complexity, stability, and system performance are interrelated. For instance, more complex ecological systems tend to be less stable, and in software engineering, software performance reliability is directly affected by software complexity. We introduce a new rigorous framework that shows how these three properties interrelate. We focus on a specific family of systems that predict input binary Markov chains. A system’s output is a binary sequence that indicates when it fails to predict the input. System complexity is the average number of information bits needed to describe the output per input bit. System stability is the discrepancy between the system’s output when presented with two random input sequences. A bound on the prediction error is derived and used as a system’s performance guarantee. It is shown that as complexity increases, stability decreases, and performance guarantee becomes more sensitive to changes in the input, making the system less robust. The analysis taken here is applicable to more general systems.
KW - concentration inequality
KW - entropy
KW - Markov chain prediction
KW - system complexity
UR - http://www.scopus.com/inward/record.url?scp=85214302441&partnerID=8YFLogxK
U2 - 10.2140/memocs.2024.12.411
DO - 10.2140/memocs.2024.12.411
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:85214302441
SN - 2326-7186
VL - 12
SP - 411
EP - 470
JO - Mathematics and Mechanics of Complex Systems
JF - Mathematics and Mechanics of Complex Systems
IS - 4
ER -