百度搜索 死在火星上 天涯 或 死在火星上 天涯在线书库 即可找到本书最新章节.
作者君在作品相关中其实已经解释过这个问题。不过仍然有人质疑——“你说得太含糊了”，“火星轨道的变化比你想象要大得多！”
那好吧，既然作者君的简单解释不够有力，那咱们就看看严肃的东西，反正这本书写到现在，嚷嚷着本书BUG一大堆，用初高中物理在书中挑刺的人也不少。
以下是文章内容：
Long-term&ions and stability of&ary orbits in our Solar system
Abstract
We&he&s of very long-term numeribsp;iions of&ary orbital motions over 109 -yr time-spans inbsp;all nine&s. A quispebsp;of our numeribsp;data shows that the&ary motion， at& in our simple dynamibsp;model，&o be&able evehis very long time-span. A bsp;look at the&-frequenbsp;oss using a lo;filter shows us&entially diffusive charabsp;of&rial&ary motion， espebsp;that of Merbsp;The behaviour of the&y of Merbsp;in our&ions is qualitatively similar to the&s from Jabsp;Laskar's sebsp;perturbation theory (e.g. emax～ 0.35 over ～± 4 Gyr). However， there are no apparent sebsp;inbsp;of&y or in in any orbital&s of the&s， whibsp;may be revealed by still&erm numeribsp;iions. ;also performed a bsp;of trial&ions inbsp;motions of&er five&s&he duration of ± 5 × 1010 yr. The& ihat&hree major resonanbsp;in the&uo system have been maintained&he 1011-yr time-span.
1 Introdu
&ion of the problem
The&ion of&ability of our Solar system has beeed over several hundred years，&he era of&on. The problem has attrabsp;many famous&ibsp;over the years and has played a&ral role in the& of non-linear dynamid bsp;theory. However， we do not& have a&e&o the&ion of&her our Solar system is stable or not. This is partly a& of the fabsp;that the&ion of&erm ‘stability’ is vague& is used iion to the problem of&ary motion in the Solar system. Absp;it is not easy to give a bsp;rigorous and physibsp;meaningful&ion of&ability of our Solar system.
Among maions of stability， here ;the Hill&ion (Gladman 1993): absp;this is not a&ion of stability， but of instability. We define a system as being unstable when a bsp;enbsp;obsp;somewhere in the system， starting from a&ain initial figuration (Chambers，&herill & Boss 1996; Ito & Tanika;1999). A system is defined as experienbsp;a bsp;enbsp;when two bodies approae another within an area of the larger Hill radius. Otherwise the system is defined as being stable. Hencefor;we&hat our&ary system is dynamibsp;stable if no bsp;enbsp;happens during the age of our Solar system， about ±5 Gyr. Ially， this&ion may be replabsp;by one in whi obsp;of any orbital bsp;betweeher of a pair of&akes&his is bebsp;we know from experiehat an orbital bsp;is very&o&o a bsp;enbsp;in&ary and protoplaary systems (Yoshinaga， Kokubo & Makino 1999). Of bsp;this& bsp;be simply applied to systems with stable orbital resonanbsp;subsp;as the&uo system.
1.2Previous studies and aims of this research
In addition to the vagueness of the& of stability， the&s in our Solar system sho;charabsp;typibsp;of dynamibsp;bsp;(Sussman & Wisdom 1988， 1992). The bsp;of this bsp;behaviour is no;uood as being a& of resonanbsp; (Murray & Holman 1999; Lebsp;Franklin & Holman 2001). However， it would require&ing over an ensemble of&ary systems inbsp;all nine&s for a period bsp;several 10 Gyr to thhly&and the long-term&ion of&ary orbits， sinbsp;bsp;dynamibsp;systems are charabsp;by their strong dependen initial s.
From that point of vie;of the previous long-term numeribsp;iions inbsp;only&er five&s (Sussman & Wisdom 1988; Kinoshita & Nakai 1996). This is bebsp;the orbital periods of&er&s are so mubsp;lohan those of the inner four&hat it is mubsp;easier to follow the system for a given&ion period. At&， the& numeribsp;iions published in journals&hose of Dunbsp;& Lissauer (1998). Although their main& ;the& of post-main-sequenbsp;solar mass loss oability of&ary orbits， they performed many&ions bsp;up to ～1011 yr of the orbital motions of the four jovian&s. The initial orbital&s and masses of&s&he same as those of our Solar system in Dunbsp;& Lissauer's paper， but they debsp;the mass of the Sun gradually in their numeribsp;experiments. This is bebsp;they bsp;the& of post-main-sequenbsp;solar mass loss in the paper.&ly， they found that the bsp;time-sbsp;of&ary orbits， whibsp;bsp;be a typibsp;indibsp;of the instability time-sbsp;is quite&ive to&e of mass debsp;of the Sun.&he mass of the Sun is&o its&&he jovian&s remain stable over 1010 yr， or perhaps longer. Dunbsp;& Lissauer also performed four similar&s on the orbital motion of seven&s (Venus to&une)， whibsp;bsp;a span of ～109 yr. Their&s on the seven&s&& prehensive， but it&hat&errestrial&s also remain stable during&egration period， maintaining almular oss.
Oher hand， in his absp;semi-analytibsp;sebsp;perturbation theory (Laskar 1988)， Laskar finds that large and irregular variations bsp;appear in the&ies and ins of&errestrial&s， espebsp;of Merbsp;and Mars on a time-sbsp;of several 109 yr (Laskar 1996). The&s of Laskar's sebsp;perturbation theory should be ed and&ed by fully numeribsp;iions.
In this paper we& preliminary&s of six long-term numeribsp;iions on all nine&ary orbits， bsp;a span of several 109 yr， and of two other&ions bsp;a span of ± 5 × 1010 yr.&al elapsed time for all&ions is&han 5 yr， using several&ed Pd workstations. One of the fual s of our long-term&ions is that Solar system&ary motioo&able in terms of the Hill stability&ioned&& over a time-span of ± 4 Gyr. Absp;in our numeribsp;iions the system ;far&able than ;is defined by the Hill stability : not only did no bsp;enbsp;happen during&egration period， but also all the&ary orbital&s have been bsp;in a narrion both in time and frequenain， though&ary motions&obsp;Sihe purpose of this paper is to& and overview the&s of our long-term numeribsp;iions， we show typibsp;example figures as evidenbsp;of the very long-term stability of Solar system&ary motion. For readers who have more spebsp;and deeper&s in our numeribsp;results， repared a ;(absp;)， where we sho;orbital&s， their lo;filtered&s， variation of Delaus and angular&um&， as of our simple time–frequenalysis on all of our&ions.
Iion 2 we briefly explain our dynamibsp;model， numeribsp;method and initial s used in our&ions.&ion 3 is&ed to a&ion of the quibsp;results of the numeribsp;iions. Very long-term stability of Solar system&ary motion is apparent both in&ary positions and orbital&s. A rough&ion of numeribsp;errors is also giveion 4 goes on to a disbsp;of the&erm variation of&ary orbits using a lo;filter and inbsp;a disbsp;of angular&um&. Iion 5， we& a& of numeribsp;iions for&er five&hat spans ± 5 × 1010 yr. Iion 6 ;disbsp;the long-term stability of the&ary motion and its possible cause.
2&ion of the numeribsp;iions
（本部分涉及比较复杂的积分计算，作者君就不贴上来了，贴上来了起点也不一定能成功显示。）
2.3 Numeribsp;method
&ilize a sed-order Wisdom–Holman symplebsp;map as our main&iohod (Wisdom & Holman 1991; Kinoshita， Yoshida & Nakai 1991) with a spebsp;start-up probsp;to redubsp;the trunbsp;error of angle variables，‘;start’(Saha & Tremaine 1992， 1994).
&epsize for the numeribsp;iions is 8 d throughout all&ions of the nine&s (N±1，2，3)， whibsp;is about 1/11 of the orbital period of the&& (Merbsp;As for&ermination of stepsize， ;follow the previous numeribsp;iion of all nine&s in Sussman & Wisdom (1988， 7.2 d) and Saha & Tremaine (1994， 225/32 d). We rouhe debsp;part of&heir stepsizes to 8 to&he stepsize a multiple of 2 in&o redubsp;the ac of round-off error in the putation probsp;Iion to this， Wisdom & Holman (1991) performed numeribsp;iions of&er five&ary orbits using the symplebsp;map with a stepsize of 400 d， 1/10.83 of the orbital period of Jupiter. Their&&o be absp;enough， whibsp;partly justifies our&hod of&ermining&epsize. However，&he&y of Jupiter (～0.05) is mubsp;smaller than that of Merbsp;(～0.2)， we need some bsp;when we pare&egrations simply in terms of stepsizes.
Iegration of&er five&s (F±)， we&he stepsize at 400 d.
;Gauss' f and g funbsp;in the symplebsp;map&her with&hird-order Halley&hod (Danby 1992) as a solver for Kepler&ions. The number of maximum&io in Halley's&hod is 15， but they never reabsp;the maximum in any of our&ions.
&erval of&a output is 200 000 d (～547 yr) for the cals of all nine&s (N±1，2，3)， and about 8000 000 d (～21 903 yr) for&egration of&er five&s (F±).
Although no output filtering ;done&he numeribsp;iions were in probsp;lied a lo;filter to the raw orbital data after ;pleted all the cals. See&ion 4.1 for&ail.
2.4 Error&ion
2.4.1&ive errors in total energy and angular&um
Absp;to one of the basibsp;properties of symplebsp;iors， whibsp;bsp;the physibsp;servative quantities well (total orbital energy and angular&um)， our long-term numeribsp;iioo have been performed with very small errors. The averaged&ive errors of total energy (～10?9) and of total angular&um (～10?11) have remained nearly bsp;throughout&egration period (Fig. 1). The spebsp;startup probsp;;start， would have redubsp;the averaged&ive error in total energy by about one order of magnitude or more.
&ive numeribsp;error of&al angular&um δA/A0 and&al energy δE/E0 in our numeribsp;iionsN± 1，2，3， where δE and δA&he absolute bsp;of&al energy and total angular&um，&ively， andE0aheir initial&he horizontal unit is Gyr.
&hat&&ing systems，&&ibsp;libraries， and& hard;architebsp;result in& numeribsp;errors， through the variations in round-off error handling and numeribsp;algorithms. In the upper panel of Fig. 1， ;rebsp;this situation in the sebsp;numeribsp;error ial angular&um， whibsp;should be rigorously preserved up to mabsp;pre.
2.4.2 Error in&ary longitudes
&he symplebsp;maps preserve total energy and total angular&um of N-body dynamibsp;systems&ly well， the degree of their&ion may not be a good measure of the absp;of numeribsp;iions， espebsp;as a measure of the positional error of&s，&he error in&ary loo&e the numeribsp;error in the&ary longitudes， we performed the following probsp;We pared the& of our main long-term&ions with&est&ions， whibsp;span mubsp;shorter periods but with mubsp;higher absp;than the main&ions. For this purpose， we performed a mubsp;more absp;iion with a stepsize of 0.125 d (1/64 of the main&ions) spanning 3 × 105 yr， starting with the same initial s as in the N?1&ion. We bsp;that this&&ion provides us with a&rue’ solution of&ary orbital&io， we pare&est&ion with the main&ion， N?1. For the period of 3 × 105 yr， we see a differen mean anomalies of the&h&he two&ions of ～0.52°(in the bsp;of the N?1&ion). This differenbsp;bsp;be&ed to the value ～8700°， about 25 rotations of&h after 5 Gyr，&he error of longitudes inbsp;linearly with time in the symplebsp;map. Similarly， the longitude error of Pluto bsp;be&ed as ～12°. This value for Pluto is mubsp;better than the& in Kinoshita & Nakai (1996)&he differenbsp;is&ed as ～60°.
3 Numeribsp;results – I. Gla the raw data
In this&ion we briefly&he long-term stability of&ary orbital motion through some snapshots of raw numeribsp;data. The orbital motion of&s indibsp;long-term stability in all of our numeribsp;iions: no orbital gs nor bsp;enbsp;between any pair of&ook place.
3.1 General&ion of&ability of&ary orbits
First， we briefly look at the general charabsp;of the long-term stability of&ary orbits. Our& here fobsp;partibsp;on the inner four&rial&s for whibsp;the orbital time-sbsp;are mubsp;shorter than those of&er five&s. As ;see bsp;from the planar orbital figurations shown in Figs 2 and 3， orbital positions of&errestrial&s differ little&he initial and final part of eaumeribsp;iion， whibsp;spans several Gyr. The solid lines&ing the& orbits of the&s lie almost within the s;of dots even in the final part of&ions (b) and (d). This ihat throughout&ire&ion period the almular variations of&ary orbital motion remain nearly the same as they&&.
&ibsp;view of the four inner&ary orbits (from the z -axis&ion) at the initial and final parts of&egrationsN±1. The axes units are au. The xy -plane is&o the invariant plane of Solar&otal angular&um.(a) The initial part ofN+1 ( t = 0 to 0.0547 × 10 9 yr).(b) The final part ofN+1 ( t = 4.9339 × 10 8 to 4.9886 × 10 9 yr).(bsp;The initial part of N?1 (t= 0 to ?0.0547 × 109 yr).(d) The final part ofN?1 ( t =?3.9180 × 10 9 to ?3.9727 × 10 9 yr). In eabsp;panel， a total of 23 684 points are plotted with an interval of about 2190 yr over 5.47 × 107 yr . Solid lines in eabsp;panel&e the& orbits of the four&rial&s (taken from DE245).
The variation of&ies and orbital ins for the inner four&s in the initial and final part of&egration N+1 is shown in Fig. 4. As&ed， the charabsp;of the variation of&ary orbital&s does not differ signifibsp;between the initial and final part of eategration， at& for Venus，&h and Mars. The&s of Merbsp;espebsp;its&y，&o bsp;to a signifibsp;extent. This is partly bebsp;the orbital time-sbsp;of the& is the shortest of all the&s， whibsp;leads to a more rapid orbital&ion than other&s; the&& may be&o instability. This& appears to be in some& with Laskar's (1994， 1996)&ions that large and irregular variations appear in the&ies and ins of Merbsp;on a time-sbsp;of several 109 yr. However， the& of the possible instability of the orbit of Merbsp;may not fatally&he global stability of the whole&ary system owing to the small mass of Merbsp;We will&ion briefly the long-term orbital&ion of Merbsp;later iion 4 using lo;filtered orbital&s.
The orbital motion of&er five&s seems rigorously stable and quite regular&his time-span (see also&ion 5).
3.2 Time–frequenbsp;maps
Although the&ary motios very long-term stability defined as the&enbsp;of bsp;enbsp;events， the bsp;nature of&ary dynamibsp;bsp;bsp;the oscillatory period and amplitude of&ary orbital motion gradually over subsp;long time-spans. Even subsp;slight flus of orbital variation in the frequenain， partibsp;in the bsp;of&h， bsp;potentially have a signifibsp;effe its surfabsp;bsp;system through solar insolation variation (bsp;Berger 1988).
To give an overview of the long-term bsp;iy in&ary orbital motion， we performed many fast Fourier transformations (FFTs) along&ime axis， and superposed the&ing periodgrams to draw two-dimensional time–frequenbsp;maps. The spebsp;approabsp;t&ime–frequenbsp;maps in this paper is very simple – mubsp;simpler than the& analysis or Laskar's (1990， 1993) frequenalysis.
Divide the lo;filtered orbital data into many fragments of the same&h. The&h of eabsp;data& should be a multiple of 2 in&o apply&.
Eabsp;fragment of&a has a large part: for example，&he ith data begins from t=ti and ends at t=ti+T， the& data& ranges from ti+δT≤ti+δT+T，&?T. We bsp;this division until we reabsp;a&ain number N by whi+T reabsp;the total&ioh.
;an FFT to eabsp;of&a fragments， and obtain n frequenbsp;diagrams.
In eabsp;frequenbsp;diagram obtained&he&h of&y bsp;be replabsp;by a grey-sbsp;(or bsp;chart.
We perform the&， and& all the grey-sbsp;(or bsp;bsp;into one graph for eategration. The horizontal axis of these ne;should&he time，&he starting times of eabsp;fragment of data (ti， where i= 1，…， n). The&ibsp;axis&s the period (or frequenbsp;of the os of orbital&s.
;adopted an FFT bebsp;of its speed，&he amount of numeribsp;data to be deposed into frequenpos is terribly huge (several tens of Gbytes).
A typibsp;example of&ime–frequenbsp;map&ed by the above probsp;is shown in a grey-sbsp;diagram as Fig. 5， whibsp;shows the variation of&y in the&y and in of&h in N+2&ion. In Fig. 5， the dark area shows that at&ime indibsp;by the value on the absbsp;the&y indibsp;by the ordinate is strohan in the lighter area around it. ;rebsp;from this map that the&y of the&y and in of&h only bsp;slightly&he&ire period bsp;by the N+2&ion. This nearly regular trend is qualitatively the same in other&ions and for other&s， although typibsp;frequenbsp;differ& by& a by&.
4.2 Long-term exbsp;of orbital energy and angular&um
We calbsp;very long-periodibsp;variation and exbsp;of&ary orbital energy and angular&um using filtered Delaus L， G， H. G and H are& to the&ary orbital angular&um and its&ibsp;po per unit mass. L is&ed to the&ary orbital energy E per unit mass as E=?μ2/2L2. If the system is pletely linear， the orbital energy and the angular&um in eabsp;frequenbsp;bin must be t. y in the&ary system bsp;bsp;an exbsp;of energy and angular&um in the frequenain. The amplitude of the&-frequenbsp;os should inbsp;if the system is unstable and breaks down gradually. However， subsp;a symptom of instability is not promi in our long-term&ions.
In Fig. 7，&al orbital energy and angular&um of the four inner&s and all nine&s are shown for&ion N+2. The&hree panels show the long-periodibsp;variation of total energy&ed asE- E0)， total angular&um ( G- G0)， and the&ibsp;po ( H- H0) of the inner four&s calbsp;from the lo;filtered Delaus.E0， G0， H0&e the initial values of eabsp;quantity. The absolute differenbsp;from the initial values is plotted in the panels. The&hree panels in eabsp;figure showE-E0，G-G0 andH-H0 of&al of nine&s. The flubsp;shown in the lower panels is virtually&irely a& of the massive jovian&s.
paring the variations of energy and angular&um of the inner four&s and all nine&s， it is apparent that the amplitudes of those of the inner&s are mubsp;smaller than those of all nine&s: the amplitudes of&er five&s are mubsp;larger than those of the inner&s. This does not&hat the&errestrial&ary subsystem is&able thaer&his is simply a& of the&ive smallness of the masses of the four&rial&s pared with those of&er jovian&s. Ahiibsp;is that the inner&ary subsystem may bee unstable more rapidly thaer one bebsp;of its shorter orbital time-sbsp;This bsp;be seen in the panels&ed asinner 4 in Fig. 7&he longer-periodid irregular oss are more apparent than in the panels&ed astotal 9. Absp;the flus in theinner 4 panels&o a&ent as a& of the orbital variation of the Merbsp;However， ;he tribution from&errestrial&s， as we will see in&&ions.
4.4 Long-term bsp;of several neighb& pairs
& us see some individual variations of&ary orbital energy and angular&um expressed by the lo;filtered Delaus. Figs 10 and 11 show long-term&ion of the orbital energy of eabsp;pla and the angular&um in N+1 and N?2&ions.&ibsp;that some&s form apparent pairs in terms of orbital energy and angular&um exbsp;In partibsp;Venus ah make a typibsp;pair. In the figures， they show&ive&ions in exbsp;of energy and positive&ions in exbsp;of angular&um. The&ive&ion in exbsp;of orbital energy means that&wo&s form a bsp;dynamibsp;system in terms of the orbital energy. The positive&ion in exbsp;of angular&um means that&wo&s are simultaneously under&ain long-term&ions. didates for&urbers are Jupiter and Saturn. Also in Fig. 11， ;see that Mars shoositive&ion in the angular&um variation to the&h system. Merbsp;exhibits&aiive&ions in the angular&um versus the&h system， whibsp;seems to be a&ion bsp;by the&ion of angular&um ierrestrial&ary subsystem.
It is not bsp;at the& why the&h pair&s a&ive&ion in energy exbsp;and a positive&ion in angular&um exbsp;ossibly explain this through the general fabsp;that there are no sebsp;terms in&ary semimajor axes up to sed-order&ion theories (bsp;Brouwer & bsp;1961; Bobsp;& Pubsp;1998). This means that the&ary orbital energy (whibsp;is&ly&ed to the semimajor axis a) might be mubsp;less&ed by&urbing&han is the angular&um exbsp;(whibsp;relates to e).&he&ies of Venus ah bsp;be disturbed easily by Jupiter and Saturn， whibsp;results in a positive&ion in the angular&um exbsp;Oher hand， the semimajor axes of Venus ah are less&o be disturbed by the jovian&s. Thus the energy exbsp;may be limited only within the&h pair， whibsp;results in a&ive&ion in the exbsp;of orbital energy in the pair.
As for&er jovian&ary subsystem， Jupiter–Saturn and Urauo make dynamibsp;pairs. However，&rength of their bsp;is not as strong pared with that of the&h pair.
5 ± 5 × 1010-yr&ions of outer&ary orbits
&he jovian&ary masses are mubsp;larger thaerrestrial&ary masses，&reat the jovian&ary system as an&&ary system in terms of&udy of its dynamibsp;stability. Henbsp;;a bsp;of trial&ions that span ± 5 × 1010 yr， inbsp;only&er five&s (the four jovian&s plus Pluto). The&s& the rigorous stability of&er&ary system&his long time-span. Orbital figurations (Fig. 12)， and variation of&ies and ins (Fig. 13) show this very long-term stability of&er five&s in both&ime and the frequenains. Although we do not sho;here，&ypibsp;frequenbsp;of the orbital os of Pluto and&her outer&s is almost bsp;during these very long-term&ion periods， whibsp;is&ed iime–frequenbsp;maps on our webpage.
Iwo&ions， the&ive numeribsp;error ial energy ;～10?6 and that of&al angular&um ;～10?10.
5.1 Resonanbsp;in the&uo system
Kinoshita & Nakai (1996)&ed&er five&ary orbits over ± 5.5 × 109 yr . They found that four major resonanbsp;betweeune and Pluto are maintained during the&egration period， and that the resonanbsp;may&he main bsp;of&ability of the orbit of Pluto. The major four resonanbsp;found in previous researbsp;are as follows. In the followiion，λ&es the mean longitude，Ω is the longitude of the asbsp;node and ? is the longitude of perihelion. Subsbsp;P and e Pluto aune.
Mean motion resoweeune and Pluto (3:2). The bsp;argument θ1= 3 λP? 2 λN??P librates around 180° with an amplitude of about 80° and a libration period of about 2 × 104 yr.
The argument of perihelion of Pluto ωP=θ2=?P?ΩP librates around 90° with a period of about 3.8 × 106 yr. The dominant periodibsp;variations of the&y and in of Pluto are synized with the libration of its argument of perihelion. This is antibsp;in the sebsp;perturbation theory strubsp;by Kozai (1962).
The longitude of the node of Pluto referred to the longitude of the node of&une，θ3=ΩP?ΩN， cirbsp;and the period of this cirbsp;is equal to the period of θ2 libration. When θ3 bees zero，&he longitudes of asbsp;nodes of&une and Pluto overlap， the in of Pluto bees maximum， the&y bees minimum and the argument of perihelion bees 90°. When θ3 bees 180°， the in of Pluto bees minimum， the&y bees maximum and the argument of perihelion bees 90° again. ;& Benson (1971) ahis type of resonanbsp;later ed by Milani， Nobili & bsp;(1989).
An argument θ4=?P??N+ 3 (ΩP?ΩN) librates around 180° with a long period，～ 5.7 × 108 yr.
In our numeribsp;iions， the resonanbsp;(i)–(iii) are well maintained， and variation of the bsp;arguments θ1，θ2，θ3 remain similar during the&egration period (Figs 14–16 ). However， the fourth resonanbsp;(iv) appears to be&: the bsp;argument θ4&es libration and cirbsp;over a 1010-yr time-sbsp;(Fig. 17). This is an&ing fabsp;that Kinoshita & Nakai's (1995， 1996) shorter&ions&&o disclose.
6 Dis
;kind of dynamibsp;mebsp;maintains this long-term stability of the&ary system? ;immediately think of t;features that may be responsible for the long-term stability. First， there&o be no signifibsp;lower-order resonanbsp;(mean motion and sebsp;between any pair among the nine&s. Jupiter and Saturn are&o a 5:2 mean motion resohe famous&&y’)， but not just in the resonanbsp;zone. Higher-order resonanbsp;may&he bsp;nature of the&ary dynamibsp;motion， but they& s as to&roy&able&ary motion within the&ime of the real Solar&he sebsp;feature， whibsp;we think is more important for the long-term stability of our&ary system， is the differen dynamibsp;distaweerial and jovian&ary subsystems (Ito & Tanika;1999， 2001). When we measure&ary&ions by&ual Hill radii (R_)，&ions amorial&s are&er than 26RH，&hose among jovian&s are&han 14RH. This differenbsp;is&ly&ed to the differeween dynamibsp;features of&rial and jovian&s.&rial&s have smaller masses， shorter orbital periods and wider dynamibsp;separation. They&rongly&urbed by jovian&hat have larger masses， longer orbital periods and narrower dynamibsp;separation. Jovian&s&&urbed by any other massive bodies.
The&errestrial&ary system is still being disturbed by the massive jovian&s. However， the wide&ion and mutual&ion among&errestrial&s&he disturbaneffebsp;the degree of disturbanbsp;by jovian&s is O(eJ)(order of magnitude of the&y of Jupiter)，&he disturbanbsp;bsp;by jovian&s is a forbsp;os having an amplitude of O(eJ).&ening of&y， for example O(eJ)～0.05， is far from suffibsp;to provoke instability ierrestrial&s having subsp;a wide&ion as 26RH. Thus we assume that the& wide dynamibsp;separation amorial&s (> 26RH) is probably one of the most signifibsp;s for maintaining&ability of the&ary system over a 109-yr time-span. Our&ailed analysis of the&ionship&ween dynamibsp;distaween&s and the instability time-sbsp;of Solar system&ary motion is now on-going.
Although our numeribsp;iions span the&ime of the Solar&he number of&ions is far from suffibsp;to fill the initial phase& is nebsp;to perform more and more numeribsp;iions to and examine iail the long-term stability of our&ary dynamics.
——以上文段引自 Ito， T.& Tanika;K. Long-term&ions and stability of&ary orbits in our Solar System. Mon. Not. R. Astron. Sobsp;336， 483–500 (2002)
这只是作者君参考的一篇文章，关于太阳系的稳定性。
还有其他论文，不过也都是英文的，相关课题的中文文献很少，那些论文下载一篇要九美元（《Nature》真是暴利），作者君写这篇文章的时候已经回家，不在检测中心，所以没有数据库的使用权，下不起，就不贴上来了。
百度搜索 死在火星上 天涯 或 死在火星上 天涯在线书库 即可找到本书最新章节.