TY - JOUR

T1 - The concept of problem complexity

AU - Salado, Alejandro

AU - Nilchiani, Roshanak

N1 - Funding Information:
This work was supported in part by the DARPA/NASA Ames Contract Number NNA11AB35C on the Fractionated Space Systems F6 project. The authors would like to thank Dr. Owen Brown for his inspiration and ideas on the contents of this paper.

PY - 2014

Y1 - 2014

N2 - Recognizing the impact of system complexity on the success of a system's development has created significant research efforts towards measuring system complexity In particular, the research community has proposed techniques to measure three types of system complexity (1) structural complexity, which measures the complexity resulting from physical interconnection of components, (2) functional complexity, which measures the complexity resulting from interconnection of system functions, and (3) organizational complexity, which measures the contractual interconnection of the different organizations developing the system The majority of these metrics focus on measuring aspects of the complexity of an existing system or design However, a metric to anticipate the complexity induced by the problem itself on a system's development is lacking We therefore present the concept of Problem Complexity as the complexity level that a set of requirements can impose to any system fulfilling them In addition, we mathematically demonstrate using the concept of joint entropy how problem complexity defines the minimum level of complexity a system can achieve for a given set of requirements The paper suggests an analytic formulation to measure the complexity induced by a set of requirements in a system's development that is based on a set of heuristics that facilitate identification of conflicts between requirements The use of such analytical formulation is showcased on a notional case-study

AB - Recognizing the impact of system complexity on the success of a system's development has created significant research efforts towards measuring system complexity In particular, the research community has proposed techniques to measure three types of system complexity (1) structural complexity, which measures the complexity resulting from physical interconnection of components, (2) functional complexity, which measures the complexity resulting from interconnection of system functions, and (3) organizational complexity, which measures the contractual interconnection of the different organizations developing the system The majority of these metrics focus on measuring aspects of the complexity of an existing system or design However, a metric to anticipate the complexity induced by the problem itself on a system's development is lacking We therefore present the concept of Problem Complexity as the complexity level that a set of requirements can impose to any system fulfilling them In addition, we mathematically demonstrate using the concept of joint entropy how problem complexity defines the minimum level of complexity a system can achieve for a given set of requirements The paper suggests an analytic formulation to measure the complexity induced by a set of requirements in a system's development that is based on a set of heuristics that facilitate identification of conflicts between requirements The use of such analytical formulation is showcased on a notional case-study

KW - Entropy

KW - Problem definition

KW - System complexity

UR - http://www.scopus.com/inward/record.url?scp=84897972186&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84897972186&partnerID=8YFLogxK

U2 - 10.1016/j.procs.2014.03.066

DO - 10.1016/j.procs.2014.03.066

M3 - Conference article

AN - SCOPUS:84897972186

SN - 1877-0509

VL - 28

SP - 539

EP - 546

JO - Procedia Computer Science

JF - Procedia Computer Science

T2 - 12th Annual Conference on SystemsEngineering Research, CSER 2014

Y2 - 21 March 2014 through 22 March 2014

ER -