-
Notifications
You must be signed in to change notification settings - Fork 5
/
Copy pathChangelog_Meigo64.txt
92 lines (51 loc) · 2.84 KB
/
Changelog_Meigo64.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
MEIGO64 - last revised 29 October 2019.
---------------------------------------
--> Update 29-October-2019 (AFV):
- Modified 'ess_local_filters' to compute elapsed time using tic-toc instead of cputime inside local searches.
--> Update 21-May-2018 (AFV):
- Modified 'ess_kernel.m' so that it computes elapsed time using tic-toc instead of cputime.
--> Update 21-Mar-2018 (YANNIK):
- Added new test problems and local solvers.
--> Update 20-Feb-2018 (JRB):
- Added IPOPT binaries for Linux/Win64
--> Update 16-May-2017 (AFV):
- Modified 'ess_kernel.m' so that it stores all the intermediate points in the optimizations,
even when no local search is used.
--> Update 02-May-2017 (AFV):
- CeSS now accepts different local methods for different threads.
- Multistart can be performed with several local methods.
- Changes in the way CeSS is invoked: it is now called similarly as other methods, i.e.:
>> MEIGO(problem,opts,'cess')
--> Update 19-April-2017 (AFV):
- eSS made robust with respect to possible crashes in local searches:
Modified "ess_local_filters" including a try-catch in the call to "ssm_localsolver".
--> Update 11-November-2016 (JRB/JEGEA):
- corrected printing of objf when all points are infeasible
- added printing fobj for x0 if x0 is supplied
--> Update 7-November-2016 (AFV):
- Provide high quality gradients to FMINCON, calculated with AMICI using adjoint sensitivities
(initially implemented by Daniel Weindl, September 2016).
--> Update 20-May-2016
- the parallel solver CeSS and CeVNS now use the Matlab Parallel Computing toolbox instead of jPar.
(the use of jPar seems to be problematic in the latest Matlab versions due to changes in the JRE).
Before using CeSS or CeVNS, launch the workers using e.g. matlabpool open local 4 for 4 threads
This is a contribution by Lukas Widmer (ETHZ).
--> Update 20-May-2016
Example:
- in the Goodwin updated example, sensisitivities are now computed with AMICI
(software from Hasenauer group)
-->Previous (9-May-2016):
What is new in MEIGO64:
- new MEIGO version for Matlab 64 bit (mex64 for most local solvers, including Win and Linux)
- includes updated mex solvers: nl2sol (for parameter estmation), misqp (for MINLP), etc.
- nl2sol accepts high-quality gradients computed by sensitivities (see Gábor and Banga, 2015)
Example:
- the Goodwin_example folder illustrates these capabilities with an oscillator model. Sensitivities
are computed with odeSD (from J. Stelling group at ETHZ), but other sensitivities solvers can be used
- this example is the same used in Gábor and Banga (2015), although no regularization here
Julio R. Banga
Email: julio@iim.csic.es
References
Gábor, A. and J. R. Banga (2015) Robust and efficient parameter estimation in dynamic
models of biological systems. BMC Systems Biology, 9:74
http://bmcsystbiol.biomedcentral.com/articles/10.1186/s12918-015-0219-2