<?xml version="1.0" encoding="UTF-8"?><!DOCTYPE article  PUBLIC "-//NLM//DTD Journal Publishing DTD v3.0 20080202//EN" "http://dtd.nlm.nih.gov/publishing/3.0/journalpublishing3.dtd"><article xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" dtd-version="3.0" xml:lang="en" article-type="research article"><front><journal-meta><journal-id journal-id-type="publisher-id">AM</journal-id><journal-title-group><journal-title>Applied Mathematics</journal-title></journal-title-group><issn pub-type="epub">2152-7385</issn><publisher><publisher-name>Scientific Research Publishing</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="doi">10.4236/am.2012.37121</article-id><article-id pub-id-type="publisher-id">AM-19863</article-id><article-categories><subj-group subj-group-type="heading"><subject>Articles</subject></subj-group><subj-group subj-group-type="Discipline-v2"><subject>Physics&amp;Mathematics</subject></subj-group></article-categories><title-group><article-title>
 
 
  An Application of Eulerian Graph to PI on &lt;i&gt;Mn&lt;/i&gt;(&lt;i&gt;C&lt;/i&gt;)
 
</article-title></title-group><contrib-group><contrib contrib-type="author" xlink:type="simple"><name name-style="western"><surname>ongfa</surname><given-names>You</given-names></name><xref ref-type="aff" rid="aff1"><sup>1</sup></xref><xref ref-type="corresp" rid="cor1"><sup>*</sup></xref></contrib><contrib contrib-type="author" xlink:type="simple"><name name-style="western"><surname>Hongyan</surname><given-names>Zhao</given-names></name><xref ref-type="aff" rid="aff1"><sup>1</sup></xref></contrib><contrib contrib-type="author" xlink:type="simple"><name name-style="western"><surname>Yijun</surname><given-names>Feng</given-names></name><xref ref-type="aff" rid="aff1"><sup>1</sup></xref></contrib><contrib contrib-type="author" xlink:type="simple"><name name-style="western"><surname>Ming</surname><given-names>Cao</given-names></name><xref ref-type="aff" rid="aff1"><sup>1</sup></xref></contrib></contrib-group><aff id="aff1"><addr-line>School of Mathematics and Computer Science, Hubei University, Wuhan P.R.China</addr-line></aff><author-notes><corresp id="cor1">* E-mail:<email>yousongfa@163.com(OY)</email>;</corresp></author-notes><pub-date pub-type="epub"><day>30</day><month>07</month><year>2012</year></pub-date><volume>03</volume><issue>07</issue><fpage>809</fpage><lpage>811</lpage><history><date date-type="received"><day>April</day>	<month>25,</month>	<year>2012</year></date><date date-type="rev-recd"><day>May</day>	<month>31,</month>	<year>2012</year>	</date><date date-type="accepted"><day>June</day>	<month>6,</month>	<year>2012</year></date></history><permissions><copyright-statement>&#169; Copyright  2014 by authors and Scientific Research Publishing Inc. </copyright-statement><copyright-year>2014</copyright-year><license><license-p>This work is licensed under the Creative Commons Attribution International License (CC BY). http://creativecommons.org/licenses/by/4.0/</license-p></license></permissions><abstract><p>
 
 
  We obtain a new class of polynomial identities on the ring of n 
  &#215; n matrices over any commutative ring with 1 by using the Swan’s graph theoretic method [1] in the proof of Amitsur-Levitzki theorem. Let be an Eulerian graph with k vertices and d edges. Further let be an integer and assume that . We prore that is an PI on 
  Mn(
  C). Standard and Chang [2] -Giambruno-Sehgal [3] polynomial identities are the spectial examples of our conclusions.
 
</p></abstract><kwd-group><kwd>Eulerian Graph; Eulerian Path; Admissible; Polynomial Identity</kwd></kwd-group></article-meta></front><body><sec id="s1"><title>1. Introduction</title><p>Let <img src="21-31556\ee4a6814-a47e-4d26-8642-72c496cb9978.jpg" /> be a finite directed graph with multiple edges allowed, and let <img src="21-31556\42526511-580d-49f7-a769-779611697f01.jpg" />denote the vertex set of <img src="21-31556\ff414f5d-6371-49c7-9a60-945b47ae5474.jpg" /> and <img src="21-31556\adb16e3a-37ad-428d-ac00-c096022b5554.jpg" /> the edges set of<img src="21-31556\b2f55dc5-ee23-46ec-abac-c4b85fc6f3a5.jpg" />. Let <img src="21-31556\082e8284-d94d-4c12-9fc1-ba06df5ef7ff.jpg" /> and <img src="21-31556\955b8fe7-cbb3-4286-a660-07350291a967.jpg" /> be the functions from <img src="21-31556\5d08efa4-5e18-4c7f-a3e8-1d6a473bdd3e.jpg" /> to <img src="21-31556\395badb0-f861-4c89-ac32-85c86a412327.jpg" /> defined by <img src="21-31556\7ce60eea-2bad-4761-a730-ef03592819cd.jpg" /> where e<sub>s</sub> is an edge from vertex i to vertex j. For a vertex <img src="21-31556\a29f540e-85f6-4349-8b75-1d04717ff54d.jpg" /> we put</p><p><img src="21-31556\664391a5-8740-41c2-8cbe-eec3051b9fe3.jpg" /></p><p>and</p><p><img src="21-31556\6f3dbd4c-2975-4e3c-a105-7f7b1abd55b4.jpg" /></p><p>We say that <img src="21-31556\0d0ea2c4-6b37-4daf-afab-f750e99e555f.jpg" /> is an Eulerian path of <img src="21-31556\d9ff5c6d-f5ec-4187-bed9-feed44a0522b.jpg" /> if <img src="21-31556\2e055498-002b-4a8e-a7f0-982847b704b7.jpg" /> is an element of Sym(d) (the symmetric group acting on the set <img src="21-31556\23aee0f1-b8fa-45e3-adce-9f96defaccb1.jpg" /> and <img src="21-31556\95925ce2-9c49-4487-b479-e74e983a4b61.jpg" /> for<img src="21-31556\092d3858-ae21-404c-b5c0-941a0cab00e0.jpg" />.</p><p>It is well known that a connected graph <img src="21-31556\d3d84cb8-0cc4-409d-a2b6-a02ba9a79fe1.jpg" /> has an Eulerin path starting at vertex p and ending at vertex q if and only if one of the following two conditions applies:</p><p>1) <img src="21-31556\ad6f0df0-689f-4406-b066-02cf34919092.jpg" />and <img src="21-31556\b575618e-4985-4eab-b400-e6a6009d5aea.jpg" /> for each<img src="21-31556\cace3827-17b8-491b-811a-3b0b115d2916.jpg" />;</p><p>2) <img src="21-31556\39bc4869-85ca-4f13-8f99-df6294176a96.jpg" />and<img src="21-31556\b0f0ea9a-5bfb-451b-ac98-6eb239223729.jpg" />, <img src="21-31556\05e80535-bf1f-4e07-a771-988496775af0.jpg" />and <img src="21-31556\f234e0a1-e8d4-462a-8850-e393fa717bf7.jpg" /> for each<img src="21-31556\9d51c6ba-3c82-47c7-a382-811e3f6c8e6b.jpg" />.</p><p>A directed connected graph <img src="21-31556\7f0d3435-7c72-4c68-bf5a-bd7c64ad4547.jpg" /> with fixed vertices p and q is called Eulerian if either condition 1) or 2) is satisfied. We note that if <img src="21-31556\16a9145b-cf94-4958-89da-871b93dda31f.jpg" /> is an Eulerian graph of type (b), then the vertices p, q are uniquely determined, but in the other case we may choose any vertex<img src="21-31556\d28700c2-962e-4a5e-8106-288b7f3a88fc.jpg" />. For an Eulerian graph <img src="21-31556\a4ac94f2-b822-40e6-aee2-c863e658fb60.jpg" /> denote by</p><p><img src="21-31556\fb829f69-c899-45d3-bc4a-1f414fbeb274.jpg" /></p></sec><sec id="s2"><title>2. Main Results</title><p>Let <img src="21-31556\714e8cd2-f8ba-4b03-bff0-ccb36e0da3e1.jpg" /> be an Eulerian graph with d edges <img src="21-31556\906ae429-6f09-4ebc-9c35-ac5b4aaf66df.jpg" /> and distinguished points p and q. the polynomial <img src="21-31556\f304b52f-e639-48ee-9e88-662c79a1c9f5.jpg" /> associated with <img src="21-31556\a49ddfa8-eaf2-4224-9d51-08e7ed485f4a.jpg" /> is defined as follows:</p><p><img src="21-31556\08feba35-946e-425a-81a0-9dd4d6ad783e.jpg" /></p><p>Thus <img src="21-31556\05957073-301a-4be3-a450-79c456173026.jpg" /> is a multilinear polynomial in the set <img src="21-31556\816deb11-fc45-4534-99b6-5a2a01cb0ec6.jpg" /> of non-commuting indeterminates.</p><p>Let <img src="21-31556\10b2fccb-45c5-4e44-ae28-51056952c2b3.jpg" /> be an integer, C a commutative ring with 1 and <img src="21-31556\f185d22e-2157-4198-839b-667c9833d199.jpg" /> a set map where the<img src="21-31556\3c7169da-5bd7-4b03-8245-0037db2e2262.jpg" />’s are the standard matrix units over C. It is clear that T can be viewed as a substitution. we shall define a directed graph <img src="21-31556\738d2848-07fe-4d28-9657-dc671c8d681e.jpg" /> induced from <img src="21-31556\81ccd9f5-fc8c-4b74-99e8-8017f1ad1c36.jpg" /> by T. First consider the directed graph on the vertex set <img src="21-31556\c724f416-7515-4c6e-8614-1efd4b8391f6.jpg" /> with edge set <img src="21-31556\42f07cd7-b7ec-4c32-95bf-6f0c90009407.jpg" /> where<img src="21-31556\3649f9b3-94f0-4e5f-8b44-091f44683760.jpg" />, <img src="21-31556\9d90c2e2-4293-4299-87c7-0902e68e77d4.jpg" /> and<img src="21-31556\5b1e0092-cb52-4249-b6be-4e7970265220.jpg" />. Now we define <img src="21-31556\c067b886-0de5-4759-852c-b269e91bd292.jpg" /> by restricting the vertex set to<img src="21-31556\677867a5-50b8-40f7-8e14-08bd67148947.jpg" />. We note that the graph so obtained need by no means be connected let alone Eulerian. If it is Eulerian however, by construction <img src="21-31556\176c1f75-de22-488f-a0ff-60e05e5f0de5.jpg" /> has at most <img src="21-31556\7d29911e-2635-4b73-8b41-06fd1717f42d.jpg" /> min <img src="21-31556\eebb1649-9897-4771-933b-f1733f040764.jpg" /> vertices, where<img src="21-31556\c395d380-f075-4989-847d-4339ba20192f.jpg" />, i.e., <img src="21-31556\56b6f2fd-52a1-426f-b5a2-a413c0a6d5fd.jpg" /> for all <img src="21-31556\46e58b54-fbbd-4e99-bd87-3bb3025ebed8.jpg" /> and <img src="21-31556\06f36891-937e-4eeb-bade-9638b746bb44.jpg" />,<img src="21-31556\1ddf3757-01c8-450c-a2e8-da5c5e202867.jpg" />. Those elements of <img src="21-31556\a68b987f-a654-49ce-b1aa-987c599f85bb.jpg" /> which do lift to an Eulerian path of <img src="21-31556\b729dfe5-baf9-4f61-9091-ee809d6c13fe.jpg" /> will be called admissible (with respect to T). It is clear that <img src="21-31556\db8c5cf1-8b84-4a3c-b72e-cb2261aecc04.jpg" /> is admissible if and only if <img src="21-31556\57a4373e-21f4-4bd0-a918-356c3e944a8d.jpg" /> is an Eulerian path of<img src="21-31556\6a536128-387c-4924-ad3d-253d2d986978.jpg" />. For the remainder of this section, we introduce Swan’s theorem and our main results.</p><p>Swan [<xref ref-type="bibr" rid="scirp.19863-ref1">1</xref>]. Let <img src="21-31556\1eb15903-c53c-4e9f-8cbf-20acfd47c135.jpg" /> be an Eulerian graph with d edges and k vertices satisfying<img src="21-31556\9fd22e32-432e-45e8-8a6f-0f9812b3c8e7.jpg" />. Then <img src="21-31556\6dbe5a47-ba0e-401f-a8ae-7d5eea00e8ac.jpg" /> has the same number of odd and even permutations (with respect to the fixed order)</p><p>Theorem 1. Let <img src="21-31556\21da8b39-9497-46cf-87e8-d9572a35621d.jpg" /> be an Eulerian graph with vertex set <img src="21-31556\ac8ee6ce-42b4-4b69-8c37-3d1b1e0df933.jpg" /> and d edges. Further let <img src="21-31556\551e2fc9-b43b-4d9c-b081-02efd62bcd0a.jpg" /> be an integer such that</p><p><img src="21-31556\e5bc1348-04f2-41e0-94c7-5aa7ff06c76d.jpg" />.</p><p>Then <img src="21-31556\17a00642-9d18-4f0c-af31-8af35827ba7b.jpg" /> is a polynomial identity on the ring <img src="21-31556\97774b4b-3489-46a8-9773-dfd39f0b5168.jpg" /> of <img src="21-31556\cc1c1b13-b9ed-4aaf-8a7a-bd6efd025502.jpg" /> matrices over a commutative ring C with 1 Corollary 2. Let <img src="21-31556\7571077f-a85b-4995-8f0a-d2fdab590e55.jpg" /> be an Eulerian graph with k vertices and d edges. Further let <img src="21-31556\20e50c54-f817-40c7-85ac-6207a1d74e29.jpg" /> be an integer and assume that<img src="21-31556\226495a5-b33d-4ddb-a3ec-c6d49dcfef01.jpg" />. Then <img src="21-31556\2d409108-d180-43cb-831f-4c3c07b811df.jpg" /> is a polynomial identity on<img src="21-31556\87dd91d6-d3ea-4658-bbc6-2f999aca9458.jpg" />.</p></sec><sec id="s3"><title>3. Proof of Theorem 1</title><p>Since <img src="21-31556\f62a579b-c0e8-4a28-929a-58ea709d0d8e.jpg" /> is multilinear, it suffices to show that <img src="21-31556\05299546-8986-4e0d-9f63-14448444c6c4.jpg" />for any substitution T of <img src="21-31556\d03cd97f-2efc-4033-95f5-f1ba4d8f4775.jpg" /> matrix units over C. Fix such an T and put<img src="21-31556\703b75aa-6d00-409b-b4e1-3285ff0d9dab.jpg" />,<img src="21-31556\4c2ea46d-b3d8-493a-8544-4fa2d504284f.jpg" />. Then</p><disp-formula id="scirp.19863-formula69420"><label>(*)</label><graphic position="anchor" xlink:href="21-31556\0d7e778e-892f-49f8-9fed-c8f11541febd.jpg"  xlink:type="simple"/></disp-formula><p>Now consider<img src="21-31556\9f8175bd-7467-4b9e-8e32-0f6d6d3b3472.jpg" />. Clearly, and summand in (*) vanishes unless, for the given<img src="21-31556\0912a4c5-faf6-451e-a69a-c4fd35dc9602.jpg" />,</p><p><img src="21-31556\2059a40a-820b-4161-bd1f-1e24078900d8.jpg" /></p><p>for all<img src="21-31556\57407d22-7671-4471-8f16-203e1c1f14d0.jpg" />, i.e., if <img src="21-31556\83856ce0-20db-42be-a799-cea15de1504f.jpg" /> is admissible. If so, on multiplying the matrix units, we obtain</p><p><img src="21-31556\a7553b50-6851-4134-8b44-7ac627a0b5ae.jpg" />.</p><p>It follows that <img src="21-31556\e468f5c3-28b1-4e3a-80ca-328f29ac88d7.jpg" />where the inner sum is taken over all admissible permutations with <img src="21-31556\67c8b1a4-923f-4848-9779-1b22a93ba7af.jpg" /> and<img src="21-31556\1ba08574-b261-40b3-bd10-dca874171d27.jpg" />. If no such admissible <img src="21-31556\5fc11de2-0232-484d-81c9-de1ca68b9ba4.jpg" /> exists, the inner sum is 0 by definition. We want to prove that this inner sum is 0 anyway. It is readily seen that for any choice of u and b, a sum and <img src="21-31556\895560c8-eaad-4ec4-93ed-d10054ad05f6.jpg" /> in the inner sum arises precisely of <img src="21-31556\bc276c02-98aa-49cf-b09d-85c99bb4c57f.jpg" /> lifts to an Eulerian path of <img src="21-31556\031be8eb-254c-46ec-9cd6-12c07551d15b.jpg" /> from (p, (u) to (q, v). Thus, on applying Swan’s theorem to <img src="21-31556\3883659e-b20d-4730-8a6a-7dfb478c176d.jpg" /> with <img src="21-31556\98c19df8-aec6-493c-bc1b-3cf2178b2656.jpg" /> and<img src="21-31556\03d17017-8975-43ad-91c8-c9cb5bc51699.jpg" />, we find that the number of even and odd admissible permutations <img src="21-31556\82d17339-f088-475f-8b0e-e5dd8522a18e.jpg" /> with <img src="21-31556\046a1a3d-1356-4202-a3fa-65a3f82d3fec.jpg" />and <img src="21-31556\1f8d5be3-5b68-4bbc-84c0-e379dd76afc6.jpg" /> coincide whence the inner sum is 0 for any choice of u and v. This completes the proof.</p></sec><sec id="s4"><title>4. Applications</title><p>1) Let <img src="21-31556\b6523a09-9476-4128-aecf-0fae4ea20a49.jpg" /> be the Eulerian graph on one vertex with d loops. Then <img src="21-31556\21b02f1b-210b-4c46-92f4-740e6c9bfd6f.jpg" /> and</p><p><img src="21-31556\8f55066f-ae80-4f01-bc12-eff38dca3332.jpg" /></p><p>the standard polynomial [<xref ref-type="bibr" rid="scirp.19863-ref2">2</xref>] in d indeterminates.</p><p><img src="21-31556.files/image002.gif" />More generally, let <img src="21-31556\e538da51-3f0e-44a4-8a62-814f627f1670.jpg" /> be the Eulerian graph on k vertices with distinguished points <img src="21-31556\f97dd24a-8cfb-4be1-b155-81370f0d4111.jpg" /> and the number <img src="21-31556\c35ca4dd-a400-48f1-b26e-67e49853baf3.jpg" /> of edges from vertex i to j:</p><p><img src="21-31556\4a532f39-6e0c-44ff-a26e-2413f42a0328.jpg" /></p><p>Now clearly <img src="21-31556\ae4c9d1d-801c-48f2-8799-1b837b449ffb.jpg" /> and</p><p><img src="21-31556\0cb810c0-11f8-48fa-b858-1d99335f035b.jpg" />k times. On putting <img src="21-31556\3e967503-891d-4460-8066-b59dff6b71f1.jpg" /> and labelling the indeterminates, corresponding to the edges from i to <img src="21-31556\1e8d7090-9c49-40df-aed0-b61e52ffe155.jpg" />by<img src="21-31556\92dd93a9-05b7-4aad-b12b-3e2ab1a6476c.jpg" />, from the corollary 2 it follows that</p><p><img src="21-31556\eb495fce-6eb1-40f7-af92-087320f514a5.jpg" /></p><p>is a polynomial identity on <img src="21-31556\3ee0a557-0602-49a2-880c-5e92e2461e67.jpg" /> [<xref ref-type="bibr" rid="scirp.19863-ref3">3</xref>] if<img src="21-31556\02e40738-8eee-41b7-a8d1-c14b7519bfee.jpg" />, i.e., if<img src="21-31556\09376600-71b1-4c89-8e39-8d7a0ad5811b.jpg" />.</p><p>2) For <img src="21-31556\6b9fc0d5-5c15-476d-8352-18341ea80b75.jpg" /> we define a sequence<img src="21-31556\96a8cfd4-0ab5-42bd-a255-f305579831e0.jpg" />, <img src="21-31556\960ed43c-d3fa-47eb-8d77-fba0278bc997.jpg" />of staircase steps, and the staircase height<img src="21-31556\0c0bae07-bc46-4905-93b6-83738edd2e1f.jpg" />. We will construct a substitution T, such that <img src="21-31556\6b77d3e2-9937-44a9-95f2-e0efc47ba600.jpg" /> lifts to the unique convering directed path of <img src="21-31556\4f2d51f7-76bc-4c1d-98e9-20da9372c7a7.jpg" /> (i.e.,<img src="21-31556\f43c3368-5826-4361-9ad8-9e55264c5509.jpg" />). First define a function <img src="21-31556\c9ae559f-12dd-46d2-a04a-048d86f4a802.jpg" /> by</p><p><img src="21-31556\c164bb76-e8e6-46b5-b109-df026d6eee42.jpg" />;<img src="21-31556\38193249-f104-41c6-baac-0fbc8fdf3562.jpg" />,</p><p><img src="21-31556\172d0e06-157b-4828-affb-822a1b1f12f3.jpg" />,<img src="21-31556\c08c6ea9-1b51-4bed-bff4-08fa41f910d5.jpg" />.</p><p>Next we define by recursion the sequence of pair<img src="21-31556\44942150-69bd-4d3e-bfd1-14ca8e63defc.jpg" />, <img src="21-31556\027b404f-ab5c-4f77-a76e-fbe8cd4662e4.jpg" />, where <img src="21-31556\ae83daf5-3b77-4317-8e26-f9be1ddd9f1a.jpg" /> is a natural number and <img src="21-31556\770758f3-55e0-4a72-9ce4-c488a49c873c.jpg" /> is a subset of<img src="21-31556\2170be3c-88ff-46bc-950c-72bfc4c674d1.jpg" />. We put <img src="21-31556\f26951e9-d9a9-4008-af22-05ff0c2d5cde.jpg" />and<img src="21-31556\9c1e1771-72ae-4e58-9a55-9c27e0441806.jpg" />. Having <img src="21-31556\0b57a469-d2e7-433b-a0b1-af1e9133e863.jpg" /> in hand<img src="21-31556\b0902e80-df96-4198-bfe7-439ff7ceef2e.jpg" />. There are three cases to consider:</p><p>a)<img src="21-31556\5f1ea99a-fc83-4ba2-8fd1-7b75674118de.jpg" />b)<img src="21-31556\46bcfeb0-54db-43d8-9594-5d9866e2925d.jpg" /><img src="21-31556\5400e541-2427-4581-98c8-364ed721cb81.jpg" />, <img src="21-31556\5879eed1-c85e-4af4-81e8-25123e1ac34c.jpg" />c)<img src="21-31556\e282f878-ab99-4c89-ae14-830a5f332b64.jpg" /><img src="21-31556\93221df1-be67-426d-ad28-d97ab382cda2.jpg" />,<img src="21-31556\dd1506fa-1968-409e-b9fc-d4b31769d4f8.jpg" />.</p><p>We now put</p><p><img src="21-31556\0cee47b6-6bf7-4ed4-8d67-edee5e9dc9a9.jpg" /></p><p>and</p><p><img src="21-31556\13bc63d2-0daa-4267-bbb2-18a327628bc8.jpg" /></p><p>Let <img src="21-31556\e6fd216f-562c-46d0-a0d7-501c37c607df.jpg" /> for all<img src="21-31556\9be368ce-d48e-44db-a51a-ec62c4003671.jpg" />, it is clear that <img src="21-31556\58ab72cd-81af-4776-87f0-9fa14347e261.jpg" /> gives a substitution of <img src="21-31556\57659501-ea15-4a1e-a589-10c1339002a7.jpg" /> matrix units over C. Now</p><p><img src="21-31556\0878a7a1-ed71-4324-88c7-dc57d8976902.jpg" /></p><p>is the unique covering directed path of <img src="21-31556\6e429994-5c4c-401e-a956-b7c2e77cd534.jpg" /> from</p><p><img src="21-31556\6f5fd671-e5e0-42d6-a54c-73c9a8dad833.jpg" />to <img src="21-31556\f55b9cb9-f7f1-4eb8-8938-6a97360edd4a.jpg" /> [4-7]. Since the <img src="21-31556\bda5d450-d203-4275-9d6a-2e3a9c14af5c.jpg" /> entry of the <img src="21-31556\42c988db-e5c6-4dc1-bc7b-55b981182c35.jpg" /> matrix <img src="21-31556\2a30369a-2edc-4d59-b236-d0cbf0779743.jpg" /> is<img src="21-31556\108d11d2-86a5-4398-9150-bee30b42ffdb.jpg" />, we have Theorem 3. Let <img src="21-31556\5c30c39e-ed46-4c40-bc03-16426c6f6288.jpg" /> be an Eulerian graph and<img src="21-31556\b5a2af79-811b-40cf-a8b5-bae3b2613c59.jpg" />. If<img src="21-31556\b69c4285-a958-43cd-a27a-10d68f0fb437.jpg" />, then <img src="21-31556\05683728-6ad7-4775-bbb1-f850e600ac6c.jpg" /> is not a polynomial identity on the ring <img src="21-31556\95f1a3d8-1fa5-4f56-a5b7-806ada7fb764.jpg" /> of <img src="21-31556\b845a667-6937-4bcc-a058-51e08eca8cb7.jpg" /> matrices over a commutative ring C with 1.</p><p>Remark. It is an obvious consequence of the above theorem that if <img src="21-31556\5a9ec5be-cd1f-4bed-b3b0-0e75064bdef1.jpg" /> is not the least integer <img src="21-31556\3a1949d8-eeec-4767-ac03-4f46f96e7712.jpg" /> for which <img src="21-31556\7206522b-23cb-4ace-a38b-e751903fc5ea.jpg" /> is not a polynomial identity on<img src="21-31556\0c99b7b1-138a-4164-8f7b-561a77a8d6ef.jpg" />.</p><p>We note that, in general min<img src="21-31556\7d28bff2-8d33-4468-8808-deeac89d048f.jpg" /> is not the least integer <img src="21-31556\7741ae83-714c-4d19-ba72-358c3ff9c683.jpg" /> for which <img src="21-31556\50581ad1-d928-47b3-b2dd-bb9ca8101d05.jpg" /> is not a polynomial identity on<img src="21-31556\0a671c56-3127-432b-a885-b533f2cb8c21.jpg" />.</p><p>Let <img src="21-31556\31883fc4-821b-4206-a052-8c7f27ddaa4a.jpg" /> be the Eulerian graph on one vertex d loops. It is easily see that</p><p><img src="21-31556\13221a01-87cb-4247-a962-dadda054ba4b.jpg" /></p><p>Thus <img src="21-31556\90fd75c0-1aab-44f7-a6d7-87d3e0c45477.jpg" /> for all <img src="21-31556\86f2f56e-96aa-4e96-9b43-f153c3b08b62.jpg" /> and the minimality assertion of the Amitsur-Levitzki theorem follows; the main part is an immediate consequence of the corollary.</p><p>Let <img src="21-31556\a410a766-3a3a-4fe5-8854-729788bc2c83.jpg" /> be the Eulerian graph on k vertices with distinguished points <img src="21-31556\b587e57e-8028-469d-a311-dd001b06591e.jpg" /> and the number <img src="21-31556\6012347e-bdb7-4b58-bf28-50e0dfaa9704.jpg" /> of edges from vertex i to j:</p><p><img src="21-31556\cf505e2a-2713-41cb-9313-1ab953cf9469.jpg" /></p><p>Analogously, for any <img src="21-31556\0bac03c8-3919-43f0-9145-d1c1056e9828.jpg" /> we have</p><p><img src="21-31556\52e09346-a241-47ac-9f32-92cc49ac5fcd.jpg" /></p><p>In consequence <img src="21-31556\c5176fc0-919a-4b3d-a7a2-9c84e6d43275.jpg" /> for all<img src="21-31556\0d0bf40b-6a88-4af9-b60b-b4c5e1c395db.jpg" />.</p><p>For <img src="21-31556\83b3e5b6-92eb-4a1d-ac73-10e3b2933966.jpg" /> we get the double Capelli polynomial; it is known, however, that in this case <img src="21-31556\969b9238-6505-4978-9811-338eb7f21cda.jpg" /> is not the smallest n for which <img src="21-31556\d24ea536-17e9-4ff1-9b26-7dabdb9f2fbc.jpg" /> is not a polynomial identity on<img src="21-31556\6bc3a9b6-efc0-4da9-99bc-b9e45168f446.jpg" />.</p><p>When <img src="21-31556\635cc26a-9fcb-4e3b-aa6f-ff519f8104f0.jpg" /> we use x, y and z instead of the symbols<img src="21-31556\52377981-9aae-49be-9707-bb4749ee7707.jpg" />, <img src="21-31556\b1e9b74c-a3b9-424b-8f9b-1b523eb9097a.jpg" />, and <img src="21-31556\a50b0637-999f-4d9e-b3c8-0e44c72a53ce.jpg" /> respectively to denote the indeterminates of the triple Capelli polynomial and continue to write m for the number of edges from vertex i to i + 1. Thus the triple Capelli polynomial is</p><p><img src="21-31556\1c578029-f2d6-4439-bec7-2a4152b73c8f.jpg" /></p></sec><sec id="s5"><title>REFERENCES</title></sec></body><back><ref-list><title>References</title><ref id="scirp.19863-ref1"><label>1</label><mixed-citation publication-type="other" xlink:type="simple">R G. Swan. “An Application of Graph Theory to Algebra”, Proc. Amer. Math. Soc. Vol.14,1963, pp.367-373. Correction, Proc. Amer. Soc. Vol.21,1969, pp.379-380.</mixed-citation></ref><ref id="scirp.19863-ref2"><label>2</label><mixed-citation publication-type="other" xlink:type="simple">Q. Chang. “Some Consequences of the Standard Polynomial”. Proc. Amer. Math. Soc. Vol.104,1988, pp.707-710.</mixed-citation></ref><ref id="scirp.19863-ref3"><label>3</label><mixed-citation publication-type="other" xlink:type="simple">A. Giambruno. S K. Sehgal. “On a Polynomial Idntity for n?n Matrices”. J.Algebra, Vol.126,1989, pp.451-453.</mixed-citation></ref><ref id="scirp.19863-ref4"><label>4</label><mixed-citation publication-type="other" xlink:type="simple">S.F.You, Y.M.Zheng and D.G.Hu.“Eulerian Graph and Polynomial Identities on Matrix Rings”. Advances in Math. Vol.32,2003, pp.425-428</mixed-citation></ref><ref id="scirp.19863-ref5"><label>5</label><mixed-citation publication-type="other" xlink:type="simple">S.F.You. “The Primitivity of Extended Centroid Extension on Prime GPI-rings”. Advances in Math. Vol.29,2000, pp.331-336.</mixed-citation></ref><ref id="scirp.19863-ref6"><label>6</label><mixed-citation publication-type="other" xlink:type="simple">S.F.You. “The Essential (one-sided) Ideal of Semiprime PI-Rings”. Acta. Math. Sinica. Vol.44,2001, pp.747-752.</mixed-citation></ref><ref id="scirp.19863-ref7"><label>7</label><mixed-citation publication-type="other" xlink:type="simple">S.F.You, M.Cao and Y.J.Feng, “Semiautomata and Near Rings”, Quantitative Logic and Soft Computing 5, World Scientific, 2012, pp.428-431. </mixed-citation></ref></ref-list></back></article>