Detalle de: saCeSS2 - global mixed-integer optimization library version 2017A

saCeSS2 - global mixed-integer optimization library version 2017A ---------------------------------------------------------------------------- saCeSS2 stands for Self-adaptive Cooperative enhanced Scatter Search 2. saCeSS2 is a new version of the saCeSS library intended for the solution of large-scale mixed-integer dynamic optimization problems. saCeSS2 incorporates several novel key features: (1) Improved asynchronous cooperation between parallel processes, especially tailored for mixed-integer problems. (2) Coarse and fine-grained parallelism. (3) Improved self-tuning strategies, suitable for large mixed-integer nonlinear optimization (new). (4) Efficient mixed-integer nonlinear optimization local solvers (new). The saCeSS2 code has been implemented using Fortran 90 and C. Parallelization has been implemented using MPI and openMP. It has been tested in Linux clusters running CentOS 6.7 and Debian 8. The saCeSS2 library also allows the solution of non-linear programming (NLP) problems, but it has specifically developed with large mixed-integer non-linear optimization problems in mind. We illustrate its performance providing the code for a set of very challenging case studies from the domain of computational biology (reverse engineering of cell signaling). These reverse engineering problems are formulated as MIDOs using the framework of logic-based ordinary differential equations: (i) SSP: the simpler case study, considering a synthetic signaling pathway with 84 continuous and 34 binary decision variables. (ii) HePG2: a second case study considering the dynamic modeling of liver cancer data, with 135 continuous and 109 binaries decision variables. (iii) HPN-DREAM: an extremely difficult case study elated with breast cancer, involving 690 continuous and 138 binary decision variables. REFERENCES D. R. Penas, D. Henriques, P. Gonzalez, R. Doallo, J. Saez-Rodriguez and J.R. Banga A parallel metaheuristic for large mixed-integer nonlinear dynamic optimization problems, with applications in computational biology. PLOS ONE, 12(8): e0182186, 2017. REQUIREMENTS saCeSS2 has the following requiriments: (1) Linux command line systems (saCeSS2 has been tested under CentOS 6.6, CentOS 6.7 and Debian 8). (2) GNU or Intel compilers (saCeSS2 has been compiled with gcc-4.4.7 and with Intel compilers: intel-13.1.1 and intel-14.0.2). (3) openMPI or IntelMPI implementation of message passing interface (MPI) (saCeSS2 has been compiled with openmpi-1.8.3, openmpi-1.8.4 and intelmpi-5.0). (4) R programming language should be installed in the system (saCeSS2 has beed tested with R-3.2.0). INSTALLATION To install saCeSS2, please perform the following steps: (1) modify the Makefile: Modify the makefile to indicate the correct paths for compilers, R, other libraries, etc. (2) clean all object files of previous installations by doing: $ make veryclean (3) create the executable by doing: $ make all (4) check if the executable named paralleltestbed has been created: $ ls [SACESS2ROOT]/bin Then, a set of scripts to reproduce results shown in the manuscript are located in the folder: reproducibility_scripts