<HTML><BODY><DIV style='font-family: "Verdana"; font-size: 10pt;'><DIV>
<DIV>First of all, the description of A073675 should be changed, replacing "divisor" with "proper divisor" and "multiple" with "proper multiple".</DIV>
<DIV> </DIV>
<DIV>Second, this is closely related to A035263: if A035263(n) = 1, A073675(n) = 2n; otherwise A073675(n)=n/2. And yes, this matches both the definition of A073675 and Clark's definition (with r=2), so those define the same sequence.</DIV>
<DIV> </DIV>
<DIV>Clark's definition will give a permutation of the natural numbers for any r >= 1. If n has not appeared in the sequence previously, it will appear at ceiling(n*r). Conversely, floor(n*r) is greater floor(m*r) for any m<n, and so greater than any preceding member of the sequence; thus no number can appear twice.</DIV>
<DIV> </DIV>
<DIV>If 0 < r < 1, the result will be a Beatty sequence for 1/r, which will thus not include all natural numbers.</DIV>
<DIV> </DIV>
<DIV>As for cycles, the answer in general is no. Taking r = phi^2, starting with 1, we get</DIV>
<DIV>1,2,5,13,34,89,... (A001519, a bisection of the Fibonacci sequence).</DIV>
<DIV> </DIV>
<DIV>Probabilistically, let probability p = P(b(n) = [n/r]) = P([n/r] > [(n-1)/r] AND f([n/r] != [[n/r]/r])) =</DIV>
<DIV>(waving our hands here, there is no guarantee these conditions are independent) 1/r * (1-p). Solving for p, we get p = 1/(1+r). Then the expected value E(ln(b(n))) = 1/(1+r)*log(n/r) + 1/(1+r)*log(n*r) = log(n) + (r-1)/(r+1)*ln(r), (more hand-waving here), so the trajectory from a given starting point can be expected to increase without limit. (Use of the log is appropriate here, since we are multiplying and dividing to get the next term.) Nothing definitive here, but I suspect that for a randomly chosen r (i.e., with probability 1), the values whose trajectory is unbounded has density 1.</DIV>
<DIV> </DIV>
<DIV>On the other hand, for the particular case r = sqrt(2), the trajectories may all be cycles. The numbers 23*2^k seem to start cycles of roughly double length at each step (verified through k=4, k=5 is next value not yet in a cycle). (Probably this is actually ceiling(C*2^k) for some real number C, since the preceding cycles start at 1,2,3,6, and 12. If so, 0.71777 < C <= 0.71875.)</DIV> <BR>-----Original Message-----<BR>From: Kimberling, Clark <A href="mailto:ck6@evansville.edu">ck6@evansville.edu</A><BR><BR>
<STYLE>
.AOLPlainTextBody {
margin: 0px;
font-family: Tahoma, Verdana, Arial, Sans-Serif;
font-size: 12px;
color: #000;
background-color: #fff;
}
.AOLPlainTextBody pre {
font-size: 9pt;
}
.AOLInlineAttachment {
margin: 10px;
}
.AOLAttachmentHeader {
border-bottom: 2px solid #E9EAEB;
background: #F9F9F9;
}
.AOLAttachmentHeader .Title {
font: 11px Tahoma;
font-weight: bold;
color: #666666;
background: #E9EAEB;
padding: 3px 0px 1px 10px;
}
.AOLAttachmentHeader .FieldLabel {
font: 11px Tahoma;
font-weight: bold;
color: #666666;
padding: 1px 10px 1px 9px;
}
.AOLAttachmentHeader .FieldValue {
font: 11px Tahoma;
color: #333333;
}
</STYLE>
<DIV id=AOLMsgPart_2_4071a9ef-a828-4610-b0c0-32b06ac427ff>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>Sequence Fans,</SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006></SPAN></FONT> </DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>A073675 starts out with 2,1,6,8,10,3,14,... , the "Rearrangement of natural numbers such that a(k) is the smallest divisor of k not included earlier and if no such divisor exists then the smallest multiple of k not included earlier."</SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006></SPAN></FONT> </DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>For any positive real r, define a sequence b(n) by</SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006></SPAN></FONT> </DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>b(n) = Floor(n/r) if this number isn't yet in the sequence; else</SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>b(n) = Floor(r*n) otherwise.</SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006></SPAN></FONT> </DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>If r=2, the sequence b also starts out 2,1,6,8,10,3,14,... Is this also A073675?</SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006></SPAN></FONT> </DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>Continuing with r=2, let c be the sequence of numbers k such that k>b(k). Is c = A036554?</SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006></SPAN></FONT> </DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>Let r = sqrt(2), for which the sequence b starts with 1,2,4,5,3,8,9,11,6,7,15,...</SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>Is this a permutation of the natural numbers? (For what choices of r is this true?)</SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>Are the terms in this sequence generated in cycles, such as 3 -> 4 ->5 ->3 </SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>and 6 -> 8 -> 11 -> 15 -> 10 -> 7 -> 9 -> 6? What can be said about the lengths of cycles?</SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006></SPAN></FONT> </DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006>Clark Kimberling </SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006> </SPAN></FONT></DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006></SPAN></FONT> </DIV>
<DIV><FONT face=Arial size=2><SPAN class=442515814-02022006></SPAN></FONT> </DIV></DIV><!-- end of AOLMsgPart_2_4071a9ef-a828-4610-b0c0-32b06ac427ff --></DIV></DIV>
<hr style="MARGIN-TOP:10px" >
<b>Try the New Netscape Mail Today!</b><br />
Virtually Spam-Free | More Storage | Import Your Contact List<br /><a href="http://mail.netscape.com">http://mail.netscape.com</a>
</BODY></HTML>