1 Introduction

This report gives the result of running the computer algebra independent integration problems (Lite version) obtained from from Albert Rich Rubi web site.

The following versions of Rubi were tested at this time. All on windows 7.

  1. Version 4.16.0.4
  2. Version 4.15.2
  3. Version 4.14.1
  4. Version 4.13.3
  5. Version 4.12.1
  6. Version 4.11
  7. Version 4.10.1
  8. Version 4.9
  9. Version 4.8
  10. Version 4.7
  11. Version 4.5
  12. Version 4.2
  13. Version 3

The command AboluteTiming[] was used to obtain the CPU time.

A time limit of 3 minutes is used for all integrals in each CAS. If the integration does not complete within this time limit then the integral is considered to have failed.

The table below gives additional break down of the grading of quality of the antiderivatives generated by each CAS. The grading is given using the letters A,B,C and F with A being the best quality. The grading is accomplished by comparing the antiderivative generated with the optimal antiderivatives included in the test suite. The following table describes the meaning of these grades.




grade

description





A

Integral was solved and antiderivative is optimal in quality and leaf size.



B

Integral was solved and antiderivative is optimal in quality but leaf size is larger than twice the optimal antiderivatives leaf size.



C

Integral was solved and antiderivative is non-optimal in quality. This can be due to one or more of the following reasons

  1. antiderivative contains a hypergeometric function and the optimal antiderivative does not.
  2. antiderivative contains a special function and the optimal antiderivative does not.
  3. antiderivative contains the imaginary unit and the optimal antiderivative does not.



F

Integral was not solved. Either the integral was returned unevaluated within the time limit, or it timed out, or CAS hanged or crashed or an exception was raised.




Based on the above, the following tables summarizes the grading for each test suite for each version

This table shows the percentage and count of solved integrals for each version. There are a total of [ 14944 ] integrals in the tests suite.






Version percentage solved number solved number failed








4.16.0.4 95.55 14279 665




4.15.2 95.537 14277 667




4.14.1 95.416 14259 685




4.13.3 95.537 14254 690




4.12.1 95.416 14207 685




4.11 95.383 14198 690




4.10.1 95.068 14172 746




4.9 95.008 14121 772




4.8 94.834 14018 823




4.7 94.493 14012 926




4.5 93.804 13850 932




4.2 93.763 13389 1094




3 92.679 11920 3024





TableĀ 1: Solved percentage over versions

This figure shows the percentage of passed integrals in each version.


pict


This Plot shows the number of A graded result for each version.


pict


This table shows the break down of different grading for each version.







Version %A %B %C %F










4.16.0.4 95.189 (14225) 0.281 (42) 0.08 (12) 4.45 (665)





4.15.2 95.169 (14222) 0.288 (43) 0.08 (12) 4.463 (667)





4.14.1 95.102 (14212) 0.234 (35) 0.08 (12) 4.584 (685)





4.13.3 95.122 (14215) 0.187 (28) 0.074 (11) 4.617 (690)





4.12.1 94.694 (14151) 0.241 (36) 0.134 (20) 4.932 (737)





4.11 94.56 (14131) 0.274 (41) 0.174 (26) 4.992 (746)





4.10.1 93.804 (14018) 0.355 (53) 0.676 (101) 5.166 (772)





4.9 91.957 (13742) 0.428 (64) 2.108 (315) 5.507 (823)





4.8 90.645 (13546) 0.917 (137) 2.242 (335) 6.196 (926)





4.7 90.471 (13520) 1.191 (178) 2.101 (314) 6.237 (932)





4.5 88.41 (13212) 1.773 (265) 2.496 (373) 7.321 (1094)





4.2 84.442 (12619) 84.442 (296) 3.172 (474) 10.406 (1555)





3 70.885 (10593) 7.796 (1165) 1.084 (162) 20.236 (3024)






TableĀ 2: Performance grading summary table over versions

This figure show the normalized mean leaf size for each version. This was normalized to the size of the optimal result.


pict


This figure show the mean leaf size for each version.


pict


This figure show the median leaf size for each version.


pict


This figure show the mean CPU time (sec) for each version.


pict