Reasoning with optional and preferred requirements
Download as PPTX, PDF0 likes537 views
This document proposes a method for reasoning with optional and preferred requirements when finding solutions to requirements problems. It represents requirements using a goal model that distinguishes between mandatory and optional goals. It presents an algorithm to efficiently find solutions that satisfy as many optional goals as possible while respecting any user-expressed preferences among the goals. The algorithm identifies all admissible solutions, applies preferences to filter and identify preferred solutions, and uses local search methods like Tabu search to optimize sets of optional goals at a large scale. An implementation and case study on enterprise portals are also discussed.
Reasoning with optional and preferred requirements
- 1. Reasoning with optional and preferred requirementsNeil A. Ernst, John Mylopoulos, Alex Borgida, Ivan JuretaUniversities of Toronto, Trento, Rutgers, Namurnernst@cs.toronto.edu
- 2. 2MotivationMost RE processes assume we derive a single implementation RE models express a landscape of alternative solutions, from which we can select those suitableQ1. how do we find solutions (efficiently)?Q2. what solution(s) should we choose?
- 3. 3The requirements problemFind a specification S which will create a system to satisfy desired requirements R in a domain D.↪ D, S ⊢ R1Let’s assume R is represented using goals.Then a solution to our problem is a set of input goals satisfying our mandatory goals. Represent this model using the Sebastiani formulation (QGM).21P. Zave and M. Jackson. Four Dark Corners of Requirements Engineering. ToSEM, 6:1–30, 1997.2Sebastiani, R., Giorgini, P., Mylopoulos, J.: Simple and Minimum-Cost Satisfiability for Goal Models.CAISE 2004
- 4. MMALORORORORBCJKP– –OR++ORFDE++– –ANDANDGH 1. Our solution must have {A,L}2. {E} would be nice, and we prefer {C} to {B} 3. S1={D,J} S2 = {D,K} S3 = {E,K} S4 = {G,H,J} 4. Preferred soln: {E,K}
- 5. 5Accounting for tastePreferences are binary relations over goals (prefer G1 G2) prefer solutions with G1 over those with G2.Options are requirements it would be nice to have in our solution(prefer O1∈S O1∉S)Noriaki Kano1 would call these attractive requirements, as separate from must-be or performance requirements.1Kano, N., et al. “Attractive quality and must-be quality.” Journal of the Japanese Society for Quality Control 14, no. 2 (1984): 39-48.
- 7. 7Finding solutionsIdentify admissible models (‘possible solutions’);Satisfy as many optional requirements as possible;Identify alternatives; Filter the set of alternatives using user-expressed preferences.
- 8. 8Finding solutions (2)Separate model into must-be and optional.Find admissible solutions:one which satisfies all must-be requirements.Find maximal set(s) of options which keep model admissible
- 9. 9Finding solutions (3)Identify alternatives (OR) Admissible solutions + optionsBoolean satisfiability solver to find #SATSet of optionset + alternativesPrune alternatives using preferencesDominate(P,Q)⇔for everyq∈Q, there is anp∈Pwhich is preferred or equivalentE.g.P = (A,B,C), Q = (B,C,D), (pref A D) P is dominant
- 10. 10Initial optimizationsScale option selection using cardinality or membership queries Filter using preferences over options Find near-optimal sets with Tabu search11Glover, Fred. Tabu Search--Part I. INFORMS Journal on Computing 1, no. 3 (January 1989): 190-206.
- 11. 11Tabu search of optionsSearch the lattice formed by subset closure of option setsImprovements measured using cardinalityTabu search allows for intelligent backtrackingLimit number of moves for speedNo guarantees of optimality, only of termination X✓
- 13. 13Experimental evaluationImplementation code at http://github.com/neilernst/er2010
- 14. SummaryWe have used options and preferences to select between solutions to the requirements problem.We introduced several optimizations in order to find solutions efficiently.14nernst@cs.toronto.edu
- 15. Thanks!
- 17. 17OutlineMotivationThe req. problemPrefs and priorities in goal modelsKano modelCase studyAn extension of the qualitative goal modeling framework of Sebastiani et al. [?], supporting stakeholder preferences and optional requirements.Macro operators for managing the scale of the possible problem and solution spaces.Algorithms to generate and compare all possible solutions.A local search algorithm to efficiently find solutions in a complex requirements model. A case study showing these concepts working on a problem at reasonable scale.
Editor's Notes
- #3: “Correct specifications, in conjunction with appropriate domain knowledge, imply the satisfaction of the requirements.”|- == derives
- #10: \\begin{figure}\\centerline{\\xymatrix @R=4mm @C=4mm{& \\{A,B,C,D\\} & \\\\\\{A,B,C\\} \\ar[ur] & \\{A,C,D\\} \\ar[u] & \\{B,C,D\\} \\ar[ul] \\\\ %p_1 \\ar[dr]^{}\\{A,B\\} \\ar[u] \\ar[uur] & \\{B,C\\} \\ar[ul] \\ar@<1ex>[uu] \\ar[ur] & \\ldots \\\\ %\\ar[ur]^{} \\ar[d]^{}\\{A\\} \\ar[u] \\ar[ur] & \\{B\\} \\ar[u] \\ar[ul] & \\dots \\\\}} \\label{fig:ex1}\\end{figure}Tabu