24小时热门版块排行榜

>论坛更新日志 (4890)
>虫友互识 (615)
>导师招生 (550)
>文献求助 (353)
>找工作 (293)
>论文投稿 (277)
>职场人生 (245)
>休闲灌水 (184)
>考博 (176)
>考研 (109)
>硕博家园 (106)
>学术会议 (96)
>博后之家 (93)
>招聘信息布告栏 (77)
>医学 (66)
>绿色求助(高悬赏) (41)

返回列表

当前主题已经存档。

天空空

木虫 (正式写手)

应助: 0 (幼儿园)
金币: 581
散金: 6
红花: 4
帖子: 596
在线: 166.1小时
虫号: 602252
注册: 2008-09-14
性别: MM
专业: 数学物理

[交流] ORMAS(Occupation Restricted Multiple Active Space Configuration Interaction)

A generalized production function (GPF) describing the entire molecular system. Such a wave function can be constructed as an antisymmetrized product of group functions of all electron groups in the molecule.
An antisymmetrized wave function describing a group of electrons in the mean field of all other groups.
a general implementation of group function theory which can perform full optimization of a wave function which is expressed as a generalized product of group function. GVB and MCSCF can be considered as two special cases of GFT.
The so-called Occupation Restricted Multiple Active Space Configuration Interaction (ORMAS) method has been introduced in GAMESS(US) is a very powerful method, and no other quantum chemistry packages can perform!
The general philosophy of ORMAS has a certain similarity to GFT in terms of multiple groups. But the two methods are actually totally different both technically and theoretically. It would be interesting to compare the two methods. The ORMAS method involves two steps: first the active space
(a set of active orbitals) is partitioned into an unrestricted number of orbitals, and secondly occupation restrictions are specified in the form of the minimal and maximal number of electrons allowed for each group. ORMAS can be considered the most general implementation of restricted active space SCF method, thus all other variations such as all CASSCF and the 3-group RASSCF(the ordinarily RASSCF, which most other packages coded) to N-group RASSCF can be can be recovered by appropriate specifications of the restriction parameters. ORMAS efficiently eliminates ineffective configurations in CI computation while keeping the flexibility to account the electron correlations among groups. One should be aware that the flexibility for group-group electron correlation comes at the price the number of exponentially increasing of configurations - even though such an increase is considerably reduced by the restrictions.
The group function method approaches the electron correlation of large systems in a different way. It also involves two steps. The first step is similar to that of ORMAS, i.e. a molecule system can be partitioned into an unrestricted number of electron groups. The second step is quite different from ORMAS: 1) the number of electrons in each group is fixed, i.e. electrons are not hopping between different groups; 2) the wave function is expressed as the antisymmetrized product of group functions; 3) each group function is optimized in the mean field of the other groups thus the small inter-group correlation is neglected. A major advantage of this approach is that the computational cost increases only linearly with the number of groups instead of exponentially as in ORMAS. In principle, any correlation method can be used for the group function of a GF group. For instance, a GF group can be further divided into multiple subgroups and the ORMAS method can be applied to these subgroups.
GAMESS(US) is a very very powerful ab initio quantum chemstry package, has many powerful and particular methods, such the previous introduced ORMAS, and GMCQDPT(General Reference Multiconfigurational Quasi-Degenerate Perturbation Theory), which is the most general implementation of Multi-Reference Perturbation Theory! and there are many many... ...
So , if someone turely want to do research in quantum chemstry , you should chose GAMESS(US), or at least, you should consider it if other packages can not solve your problem or meet your requirements!

[ Last edited by mingdong on 2009-6-13 at 21:57 ]

回复此楼