Advertisement

Operations-Research-Spektrum

, Volume 2, Issue 4, pp 235–249 | Cite as

The user interface and performance of commercial mixed integer programming software

  • H. Peeters
OR Software

Summary

Today's mixed integer programming software packages are still not sophisticated enough to be treated as black boxes. Their efficiency can be augmented considerably by a deep knowledge of their functioning and the degrees of freedom they offer to the user. This paper describes at first the choices of the underlying branch and bound method in seven representative MIP-systems and demonstrates their usage, their profitability and their catches at illustrative examples. An account follows of the procedures and administrative facilities of commercial MIP-modules as far as they are submitted to user influence. An outlook is given on future developments, which can further improve the acceptance and the performance of MIP-software. Finally, some effort assessment criteria are proposed, which help judging the amount of work ahead before and at different stages during the solution process.

Keywords

Software Package User Interface Future Development Programming Software Integer Programming 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

List of Symbols

c

vector of objective function coefficients of continuous variables

C

matrix of coefficients of continuous variables

d

vector of objective function coefficients of integer variables

D

matrix of coefficients of integer variables

d1

distance betweenFas and X 0 *

d2

distance betweenFas andFbk

dj

estimated functional value degradation when lowering the value ofyj by one unit

dj+

estimated functional value degradation when rising the value ofyj by one unit

Ek

estimation of best functional value of any integer solution contained in waiting nodek

Fas

aspiration level; nodes with functional value worse thanFas are postponed

Fase

aspiration level; nodes with estimationEk worse thanFase are postponed

FbI

functional value of best integer solution known so far

Fbk

best functional value of all waiting nodes

FIT0

functional value of integer optimum

Fk

functional value of waiting nodek

F0

functional value of continuous optimum

M

big positiv real number

p

maximum percentage deviation ofFITO fromFbI

skj

successor nodej

u

distance by whichX 0 * is lower thanFbI

X0*

cutoff value; nodes with functional value beyondX 0 * are finally discarded

x

vector of integer variables

y

vector of integer variables

yj

integer variablej

yj(k)

value of integer variablej at nodek

wj

associated weight of integer variableyj in a Special Ordered Set

¯w

average weight of a Special Ordered Set

[⋯ ]

greatest integer smaller than or equal to the argument

Zusammenfassung

Heute verfügbare Softwarepakete der gemischt-ganzzahligen Programmierung sind noch nicht fortgeschritten genug, um als Black Boxes behandelt zu werden. Ihre Effizienz kann durch eine gründliche Kenntnis ihrer Funktionsweise und der dem Benutzer angebotenen Freiheitsgrade beträchtlich gesteigert werden. Diese Veröffentlichung beschreibt zuerst die Auswahlmöglichkeiten der zugrundeliegenden Branch und Bound Methode in sieben repräsentativen MIP-Systemen und demonstriert anhand illustrativer Beispiele deren Anwendung, deren Nutzen und deren Fußangeln. Im Anschluß werden Proceduren und administrative Funktionen in kommerziellen MIP-Modulen, soweit sie dem Benutzereinfluß unterliegen, dargestellt. Ein Ausblick auf zukünftige Entwicklungen zeigt, wie sich Akzeptanz und Leistungvon MIP-Software steigern können. Schließlich werden einige Kriterien zur Aufwandsabschätzung vor und in verschiedenen Stadien während des Lösungsprozesses vorgeschlagen.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Abadie J (ed) (1970) Integer and nonlinear programming. Elsevier, AmsterdamGoogle Scholar
  2. 2.
    Beale EML (1977) The state of the art in numerical analysis. In: Jacobs D (ed) Proceedings of the conference on the state of the art in numerical analysis. Academic Press, New York, pp 409–448Google Scholar
  3. 3.
    Beale EML (1979) Branch and bound methods for mathematical programming systems. Ann Discrete Math 5:201–219Google Scholar
  4. 4.
    Beale, EML, Forrest JJH (1976) Global optimization using special ordered sets. Math Programming 10:52–69Google Scholar
  5. 5.
    Beale EML, Forrest JJH (1978) Global optimization as an extension of integer programming. In: Dixon LCW, Szegö GP (eds) Towards global optimization. North Holland, Amsterdam, pp 131–149Google Scholar
  6. 6.
    Beale EML, Small RE (1965) Mixed integer programming by a branch and bound technique. In: Kalenich WA (ed) Proceedings of the IFIP congress, vol. 2. Macmillan, London, pp 450–451Google Scholar
  7. 7.
    Beale EML, Tomlin JA (1970) Special facilities in a general mathematical programming system for non-convex problems using ordered sets of variables. In: Lawrence J (ed) Proceedings of the fifth international conference on operations research, Venedig 1969. Travistock, London, pp 447–454Google Scholar
  8. 8.
    Beale EML, Tomlin JA (1972) An integer programming approach to a class of combinatorial problems. Math Programming 3:339–344Google Scholar
  9. 9.
    Benders JF (1962) Partitioning proceedures for solving mixed-variables programming problems. Numer Math 4: 238–252Google Scholar
  10. 10.
    Benichou M, Gauthier JM, Girodet P, Hentges G, Ribiere G, Vincent O (1971) Experiments in mixed-integer linear programming. Math Programming 1:76–94Google Scholar
  11. 11.
    Dixon LCW, Szegö GP (eds) (1978) Towards global optimization. North Holland, AmsterdamGoogle Scholar
  12. 12.
    Driebeek NJ (1966) An algorithm for the solution of mixed-integer programming problems. Manag Sci 12:576–587Google Scholar
  13. 13.
    Forrest JJH, Hirst JPH, Tomlin JA (1974) Practical solution of large mixed-integer programming problems with UMPIRE. Manag Sci 20:736–774Google Scholar
  14. 14.
    Forrest JJH, Crowder H (1979) Report of the session on current state of computer codes for discrete optimization. Ann Discrete Math 5:271–274Google Scholar
  15. 15.
    Gauthier JM, Ribière G (1977) Experiments in mixed-integer linear programming using pseudo-costs. Math Programming 12:26–47Google Scholar
  16. 16.
    Geoffrion AM (1974) Lagrange relaxation for integer programming. Math Programming Study 2:82–114Google Scholar
  17. 17.
    Geoffrion AM, Marsten RE (1972) Integer programming algorithms: A framework and state of the art survey. Manag Sci 18:465–491Google Scholar
  18. 18.
    Geoffrion AM, Nauss R (1977) Parametric and postoptimality analysis in integer linear programming. Manag Sci 23:453–466Google Scholar
  19. 19.
    Gianessi F, Miliotis P (1979) Report of the session on codes for special problems. Ann Discrete Math 5:275–278Google Scholar
  20. 20.
    Granot F, Hammer PL (1972) On the use of Boolean functions in 0–1 programming. Method OR 12:154–184Google Scholar
  21. 21.
    Johnson EL, Suhl U (1979) Experiments in integer programming. Research Report, IBM Thomas J. Watson Research Center Yorktown Heights, New YorkGoogle Scholar
  22. 22.
    Land A, Powell S (1979) Computer codes for problems of integer programming. Ann Discrete Math 5:221–269Google Scholar
  23. 23.
    Marsten RE, Morin TL (1978) A hibrid approach to discrete mathematical programming. Math Programming 14: 21–40Google Scholar
  24. 24.
    Mevert P, Suhl U (1976) Lösung gemischt-ganzzahliger Planungsprobleme. In: Noltemeier H (ed) Computergestützte Planungssysteme. Physica-Verlag, Würzburg Wien, S 128–154Google Scholar
  25. 25.
    Peeters, H.: Empirischer Leistungsvergleich der kommerziellen Softwaresysteme APEX III, FMPS, MPSX-MIP/370 bei gemischt-ganzzahligen Optimierungsproblemen. In: Schwarze, J. et al. (eds) Proceedings in Operations Research, Vol 9. Physica-Verlag, Würzburg Wien, S 551–558Google Scholar
  26. 26.
    Powell S, Miliotis P (1979) Report of the session on current computer codes. Ann Discrete Math 5:279–283Google Scholar
  27. 27.
    Shapiro JF (1979) A survey of lagrangean techniques for discrete optimization. Ann Discrete Math 5:113–138Google Scholar
  28. 28.
    Spielberg K (1979) Enumerative methods in integer programming. Ann Discrete Math 5:139–183Google Scholar
  29. 29.
    Tomlin JA (1970) Branch and bound methods for integer and non-convex programming. In: Abadie J (ed) Integer and nonlinear programming. Elsevier, Amsterdam, pp 437–450Google Scholar
  30. 30.
    Williams HP (1974) Experiments in the formulation of integer programming problems. Math Programming Study 2:180–197Google Scholar
  31. 31.
    Williams HP (1978) Model building in mathematical programming. Wiley & Sons, New YorkGoogle Scholar
  32. 32.
    APEX III Reference Manual, Control Data Corporation, version 1.2, revision G, 1979, 76070000 GGoogle Scholar
  33. 33.
    FMPS Programmer Reference, Sperry Univac, version 7R2, 1977, UP-8198Google Scholar
  34. 34.
    LP-IP 2900 Reference, International Computers Ltd., version 1, 1978, 6880/1Google Scholar
  35. 35.
    MPS Implementation Guide, Honeywell Bull, version 1, 1975, 66A3-DD63Google Scholar
  36. 36.
    MPS Mixed Integer Programming User's Guide, Honeywell Bull, version 1, 1974, 66A3-DD62Google Scholar
  37. 37.
    MPSX/370 Program Reference Manual, IBM, level 1M4, 2. ed. 1976, SH 19-10-1Google Scholar
  38. 38.
    MIP/370 Program Reference Manual, IBM, level 1M1, 2. ed. 1975, SH 19-1099-1, revised 1979 by SN 19-1132Google Scholar
  39. 39.
    SCICONIC User's Guide, Scicon Computer Services Ltd., version 4.1,1978Google Scholar
  40. 40.
    TEMPO User's Manual, Burroughs Corp., 1975, 1073665Google Scholar

Copyright information

© Springer-Verlag 1981

Authors and Affiliations

  • H. Peeters
    • 1
  1. 1.FB Ökonomie, Quantitative Methoden der BWLJohann Wolfgang Goethe-UniversitätFrankfurt/Main 1

Personalised recommendations