Values of Ducci Periods for Sequences on ℤ m n \mathbb{Z}_{m}^{n} Mark L. Lewis, Shannon M. Tefft Abstract . Let D : ℤ m n → ℤ m n D:\mathbb{Z}_{m}^{n}\to\mathbb{Z}_{m}^{n} be defined so that D ⁡ ( x 1 , x 2 , … , x n ) = ( x 1 + x 2 ​ mod ​ m , x 2 + x 3 ​ mod ​ m , … , x n + x 1 ​ mod ​ m ) . D(x_{1},x_{2},...,x_{n})=(x_{1}+x_{2}\;\text{mod}\;m,x_{2}+x_{3}\;\text{mod}\;m,...,x_{n}+x_{1}\;\text{mod}\;m). We call D D the Ducci function and the sequence { D α ​ ( 𝐮 ) } α = 0 ∞ \{D^{\alpha}(\mathbf{u})\}_{\alpha=0}^{\infty} the Ducci sequence of 𝐮 \mathbf{u} for 𝐮 ∈ ℤ m n \mathbf{u}\in\mathbb{Z}_{m}^{n} . Every Ducci sequence enters a cycle, so we can let Per ​ ( 𝐮 ) \text{Per}(\mathbf{u}) be the number of tuples in the Ducci cycle of 𝐮 \mathbf{u} , or the period of 𝐮 \mathbf{u} . In this paper, we will look at what different possible values of Per ​ ( 𝐮 ) \text{Per}(\mathbf{u}) we can have and some conditions that if 𝐮 \mathbf{u} meets at least one of them, 𝐮 \mathbf{u} will generate a period smaller than the maximum period. 1991 Mathematics Subject Classification 20D60, 11B83, 11B50 1. Introduction We begin by introducing an endomorphism on \mathbb{Z}_{m}^{n} known as the Ducci function, D , with [ 1 , 9 , 11 ] being a few examples that call it this. Namely, we define D so that D(x_{1},x_{2},...,x_{n})=(x_{1}+x_{2}\;\text{mod}\;m,x_{2}+x_{3}\;\text{mod}\;m,...,x_{n}+x_{1}\;\text{mod}\;m). We also call the sequence \{D^{\alpha}(\mathbf{u})\}_{\alpha=0}^{\infty} the Ducci sequence of \mathbf{u} for \mathbf{u}\in\mathbb{Z}_{m}^{n} . As [ 1 ] notes, every Ducci sequence on \mathbb{Z}_{m}^{n} enters a cycle because \mathbb{Z}_{m}^{n} is finite. More specifically, we define Definition 1 . The Ducci cycle of \mathbf{u} is \{\mathbf{v}\mid\exists\alpha\in\mathbb{Z}^{+}\cup\{0\},\beta\in\mathbb{Z}^{+}\ni\mathbf{v}=D^{\alpha+\beta}(\mathbf{u})=D^{\alpha}(\mathbf{u})\} . The length of \mathbf{u} , \mathbf{Len(u)} , is the smallest \alpha satisfying the equation \mathbf{v}=D^{\alpha+\beta}(\mathbf{u})=D^{\alpha}(\mathbf{u}) for some v\in\mathbb{Z}_{m}^{n} the period of \mathbf{u} , \mathbf{Per(u)} , is the smallest \beta that satisfies the equation. To demonstrate this definition, we look at the Ducci sequence of (0,0,2)\in\mathbb{Z}_{4}^{3} : (0,0,2),(0,2,2),(2,0,2),(2,2,0),(0,2,2) . Here we see that \text{Len}(0,0,2)=1 and \text{Per}(0,0,2)=3 . A particular Ducci sequence of interest is the Ducci sequence of (0,0,...,0,1) in \mathbb{Z}_{m}^{n} , known as the basic Ducci sequence . The basic Ducci sequence is first introduced by [ 9 ] on page 302 and is also used by [ 1 , 8 , 11 ] . Define \mathbf{P_{m}(n)}=\text{Per}(0,0,...,0,1) and \mathbf{L_{m}(n)}=\text{Len}(0,0,...,0,1). This notation is similar to the notation that [ 1 ] uses in Definition 5, and P_{m}(n) is also similar to the notation used by [ 9 , 11 ] . This sequence is important is because of Lemma 1 in [ 1 ] , which says that for \mathbf{u}\in\mathbb{Z}_{m}^{n} , \text{Len}(\mathbf{u})\leq L_{m}(n) and \text{Per}(\mathbf{u})|P_{m}(n) . This provides maximum values for the length and periods of tuples in \mathbb{Z}_{m}^{n} . The notation P_{m}(n) for the maximum period on \mathbb{Z}_{m}^{n} is also used by [ 2 ] on page 858. Discussing a value for a maximum period also begs the question about when tuples do or do not generate this period. In this paper, we would like to focus on some different cases where this happens. We will begin by examining a smaller period that appears when n=3 by proving the following Lemma: Lemma 2 . Suppose \mathbf{u}=(x_{1},x_{2},x_{3})\in\mathbb{Z}_{m}^{3} such that x_{1}+x_{2}+x_{3}\equiv 0\;\text{mod}\;m . Then (1) Every tuple in the Ducci sequence of \mathbf{u},\;\mathbf{v}=(x_{1}^{\prime},x_{2}^{\prime},x_{3}^{\prime}) , satisfies x_{1}^{\prime}+x_{2}^{\prime}+x_{3}^{\prime}\equiv 0\;\text{mod}\;m. (2) D^{2}(\mathbf{u})=H(\mathbf{u}) . (3) \mathbf{u} belongs to a Ducci cycle. (4) If m odd, then unless x_{1}=x_{2}=x_{3} , Per(\mathbf{u})=6 . (5) If m is even and the x_{i} are not all equal, then either \text{Per}(\mathbf{u})=6 , or \text{Per}(\mathbf{u})=3 and x_{1},x_{2},x_{3}\in\{0,\frac{m}{2}\} . where H is an endomorphism on \mathbb{Z}_{m}^{n} defined so H(x_{1},x_{2},...,x_{n})=(x_{2},x_{3},...,x_{n},x_{1}). The function H is first defined by [ 9 ] on page 302, and is also used by [ 1 , 11 ] . After this, we would like to narrow our focus to when m is prime to show that \text{Per}(\mathbf{u})=P_{m}(n) for most \mathbf{u}\in\mathbb{Z}_{m}^{3} : Theorem 3 . Let n=3 , m an odd prime, and \mathbf{u}=(x_{1},x_{2},x_{3})\in\mathbb{Z}_{m}^{3} . If x_{1}+x_{2}+x_{3}\not\equiv 0\;\text{mod}\;m and x_{1},x_{2},x_{3} not all equal, then Per(\mathbf{u})=P_{m}(3) . We will then take a look at some specific cases where n,m are both prime but n may not be 3 and what periods are possible for these cases, as well as how we were able to determine those periods. Finally, we would like to look at a specific case where n=m is prime, and determine all of the possible periods for a tuple \mathbf{u}\in\mathbb{Z}_{m}^{n} for this case: Theorem 4 . Let n=m=p be an odd prime and let \delta be the order of 2 mod p . Then there are exactly 3 distinct period lengths for \mathbf{u}\in\mathbb{Z}_{p}^{p} : (1) Per(\mathbf{u})=1 , which can only happen when \mathbf{u}=(0,0,...,0) . (2) Per(\mathbf{u})=\delta , which can only happen if \mathbf{u}=(x,x,...,x) for some x\in\mathbb{Z}_{p} . (3) Per(\mathbf{u})=P_{p}(p)=p*\delta . The work in this paper was done while the second author was a Ph.D. student at Kent State University under the advisement of the first author and will appear as part of the second author’s dissertation. 2. Background The Ducci function on \mathbb{Z}_{m}^{n} is first examined by [ 21 ] , as well as [ 1 , 2 , 8 ] . However, this endomorphism is based on another function, \bar{D} , where \bar{D}(x_{1},x_{2},...,x_{n})=(|x_{1}-x_{2}|,|x_{2}-x_{3}|,...,|x_{n}-x_{1}|) and either \bar{D}:\mathbb{Z}^{n}\to\mathbb{Z}^{n} or \bar{D}:(\mathbb{Z}^{+}\cup\{0\})^{n}\to(\mathbb{Z}^{+}\cup\{0\})^{n} . A few sources that define Ducci in this way are [ 9 , 10 , 11 , 15 ] . Other papers, like [ 4 , 6 , 20 ] , look at Ducci defined on \mathbb{R}^{n} , where we once more use the formula given for \bar{D} . For the sake of simplicity, we will refer to both Ducci on \bar{D}:\mathbb{Z}^{n}\to\mathbb{Z}^{n} and \bar{D}:(\mathbb{Z}^{+}\cup\{0\})^{n}\to(\mathbb{Z}^{+}\cup\{0\})^{n} as Ducci on \mathbb{Z}^{n} . This works because if \mathbf{u}\in\mathbb{Z}^{n} , then \bar{D}(\mathbf{u})\in(\mathbb{Z}^{+}\cup\{0\})^{n} . There are a few findings from Ducci on \mathbb{Z}^{n} that we would like to note before moving on. First, every Ducci sequence on \mathbb{Z}^{n} enters a cycle. This is discussed in [ 5 , 9 , 11 , 15 ] . Additionally, we have that all of the tuples belonging to a Ducci sequence satisfy the condition that all of their entries belong to \{0,c\} for some c\in\mathbb{Z} , which is proved in Lemma 3 of [ 15 ] . For the Ducci case on \mathbb{Z}^{n} , this makes the Ducci case on \mathbb{Z}_{2}^{n} important because D(\lambda\mathbf{u})=\lambda D(\mathbf{u}) , where we use the same definition as Ducci on \mathbb{Z}_{m}^{n} . We now return to Ducci on \mathbb{Z}_{m}^{n} and our original definition for D . We begin by presenting a few more definitions that we would like to use throughout the paper. If \mathbf{u},\mathbf{v}\in\mathbb{Z}_{m}^{n} , then \mathbf{v} is called a predecessor of \mathbf{u} if D(\mathbf{v})=\mathbf{u} . Let K(\mathbb{Z}_{m}^{n}) be the set of all tuples in \mathbb{Z}_{m}^{n} that belong to a Ducci cycle for some \mathbf{u}\in\mathbb{Z}_{m}^{n} . This definition is first introduced in Definition 4 of [ 1 ] . This Definition also states that K(\mathbb{Z}_{m}^{n}) is a subgroup of \mathbb{Z}_{m}^{n} , and a proof is in Theorem 1 of [ 12 ] . There are a couple of papers that have looked at the value of P_{m}(n) . One of these papers is [ 1 ] , which in Proposition 4 shows that if p is prime, then P_{p}(p^{k}n_{1})=p^{k}P_{p}(n_{1}) . Theorem 8.2 of [ 2 ] , shows that P_{m}(n) can be broken down in terms of P_{p}(n) where m=p^{l} for some prime p and l\geq 1 . As an example of one of these break downs, Corollary 8.3 isolates to the case where p\nmid 2n , and p^{2}\nmid(2^{p-1}-1) to say P_{p^{l}}(n)=p^{l-1}P_{p}(n) . Theorem 3.2 of [ 8 ] provides more significance to this result, by proving that if m=m_{1}m_{2} where gcf(m_{1},m_{2})=1 , then P_{m}(n)=lcm\{P_{m_{1}}(n),P_{m_{2}}(n)\} , which allows us to narrow our examination of the value of P_{m}(n) to when m is a prime power, and therefore to the formula that [ 2 ] provides in Theorem 8.2. However, in this paper, we are not as interested in the actual value of P_{m}(n) . Rather, we would like to look at the set \{\text{Per}(\mathbf{u})\;|\;\mathbf{u}\in\mathbb{Z}_{m}^{n}\} as a whole. How large is this set? For each value of \text{Per}(\mathbf{u}) in this set, how many different tuples generate that period? Can we find conditions \mathbf{u} must meet to generate that period? The one value of \text{Per}(\mathbf{u}) that is always a period for some tuple in \mathbb{Z}_{m}^{n} for every n,m is \text{Per}(0,0,...,0)=1 . We define \vec{\mathbf{0}}=(0,0,...,0) . From Remark 2 of [ 1 ] , we know that the only time \text{Per}(\mathbf{u})=1 is when the Ducci cycle of \mathbf{u} is \{\vec{\mathbf{0}}\} . If the Ducci cycle of \mathbf{u} is \{\vec{\mathbf{0}}\} , then we say that the Ducci sequence of \mathbf{u} vanishes , a term that is first used on page 117 by [ 5 ] and also by [ 1 , 10 ] . One of the most well-known facts about Ducci on \mathbb{Z}^{n} is that all Ducci sequences vanish if n is a power of 2 . The first paper to prove this is [ 7 ] , as noted by [ 3 , 6 ] . The papers [ 10 , 15 ] also credit [ 7 ] for a proof of this. Unfortunately, we are unable to find a copy of [ 7 ] to confirm this ourselves. A review of the paper can be found at [ 19 ] . Nonetheless, many other papers have proofs of this fact as well, including [ 9 , 10 , 17 , 18 ] . This can be extended to Ducci on \mathbb{Z}_{m}^{n} . If n=2^{k} and m=2^{l} , then [ 21 ] proves in (I) on page 103 that all tuples in \mathbb{Z}_{m}^{n} vanish. This is proved again by [ 8 ] . In Theorem 2 of [ 12 ] , we prove this once more, plus that L_{m}(n)=2^{k-1}(l+1) . Therefore, in this case, \{\text{Per}(\mathbf{u})\;|\;\mathbf{u}\in\mathbb{Z}_{m}^{n}\}=\{1\} . For another possible period, we look at the tuples where all of the entries are the same, nonzero value. We say a tuple satisfies the uniformity condition if all of its entries are nonzero and equal. We now prove a lemma about tuples that meet this condition: Lemma 5 . Suppose \mathbf{u}=(x,x,...,x) for some 01 where m_{1}\geq 1 is odd and l\geq 0 . (1) Assume m_{1}=1 . Then \mathbf{u} vanishes for every x\in\mathbb{Z}_{m} . (2) Assume m_{1}>1 . Let \delta be the multiplicative order of 2\;\text{mod}\;m_{1} then \text{Per}(\mathbf{u})=\delta when m_{1}\nmid x and \mathbf{u} vanishes when m_{1}|x . Proof. We first note that D^{\alpha}(x,x,...,x)=2^{\alpha}(x,x,...,x) . \mathbf{(1)} : Assume m=2^{l} . Then D^{l}(x,x,...,x)=2^{l}(x,x,...,x) , which is \vec{\mathbf{0}} . Therefore, (x,x,...,x) vanishes for all x\in\mathbb{Z}_{m} . \mathbf{(2)} : First consider when l=0 so m=m_{1}>1 is odd. Let \delta be the multiplicative order of 2\;\text{mod}\;m . Then D^{\delta}(x,x,...,x)=2^{\delta}(x,x,...,x), or (x,x,...,x) . Since \delta is the smallest value where this can happen, \text{Per}(\mathbf{u})=\delta . Now let m=2^{l}m_{1} where m_{1}>1 and l>0 . Let \delta be the multiplicative order of 2\;\text{mod}\;m_{1} . If m_{1}|x , then D^{l}(x,x,...,x)=2^{l}(x,x,...,x), which is \vec{\mathbf{0}} , so \mathbf{u} vanishes. Now assume m_{1}\nmid x and 2^{l}|x . Then D^{\delta}(x,x,...,x)=2^{\delta}(x,x,...,x). This is (x,x,...,x) . Once more, \delta is the smallest value where this happens, so \text{Per}(\mathbf{u})=\delta . Suppose now that x is odd or there exists 1\leq l_{1}L_{2}(3) . Similarly, this will happen any time L_{m}(n)>L_{m_{1}}(n) . In other cases when n is even, other tuples will appear in the transition graph when moving from \mathbb{Z}_{m_{1}}^{n} to \mathbb{Z}_{m}^{n} because every tuple with a predecessor will go from having m_{1} predecessors to m predecessors. This is because of Theorem 4 of [ 12 ] , which says if n is even, then every tuple in \mathbb{Z}_{m}^{n} that has a predecessor has exactly m predecessors. As a result, in both of these cases, these new tuples that appear also have the smaller period, despite not belonging to \mathbb{Z}_{m_{1}}^{n} . We have a similar result when n is not prime. If n=n_{1}n_{2} , (x_{1},x_{2},...,x_{n_{1}})\in\mathbb{Z}_{m}^{n_{1}} , and (x_{1},x_{2},...,x_{n_{1}},x_{1},...,x_{n_{1}},...,x_{1},...,x_{n_{1}})\in\mathbb{Z}_{m}^{n} where (x_{1},x_{2},...,x_{n_{1}}) is repeated n_{2} times, then \text{Per}(x_{1},x_{2},...,x_{n_{1}},x_{1},...,x_{n_{1}},...,x_{1},...,x_{n_{1}})=\text{Per}(x_{1},x_{2},...,x_{n_{1}}). This is because D(x_{1},x_{2},...,x_{n_{1}},x_{1},...,x_{n_{1}},...,x_{1},...,x_{n_{1}}) =(D(x_{1},x_{2},...,x_{n_{1}}),D(x_{1},...,x_{n_{1}}),...,D(x_{1},...,x_{n_{1}})). There is one last possible type of period that we would like to discuss. Consider a tuple, (x_{1},x_{2},...,x_{n}) . We say it satisfies the sum condition if x_{1}+x_{2}+\cdots+x_{n}\equiv 0\;\text{mod}\;m. For some n,m , all tuples satisfying the sum condition generate a smaller period than P_{m}(n) . If n=3 , tuples satisfying the sum conditions always have a period of 6 , as we wish to prove in Lemma 2 . However, there are n,m where this does not happen. For example, proving Theorem 4 will show this is not the case when n=m is prime. We will examine this topic more for when n\neq 3 where both n,m are prime in Section 5 . Before moving on, we would like to note that for certain n,m , there are other tuples \mathbf{u}\in\mathbb{Z}_{m}^{n} such that \text{Per}(\mathbf{u})t\geq 1 , a_{r+t}=a_{t}a_{r}+b_{t}c_{r}+c_{t}b_{r} b_{r+t}=a_{t}b_{r}+b_{t}a_{r}+c_{t}c_{r} c_{r+t}=a_{t}c_{r}+b_{t}b_{r}+c_{t}a_{r}. When n=3 , a_{r},b_{r}, and c_{r} are all very close in value, which we can clarify in the following lemma: Lemma 8 . Let n=3 . Then • If r\equiv 0\;\text{mod}\;6 , a_{r}=b_{r}+1=c_{r}+1 . • If r\equiv 1\;\text{mod}\;6 , then c_{r}=a_{r}-1=b_{r}-1 . • If r\equiv 2\;\text{mod}\;6 , then b_{r}=a_{r}+1=c_{r}+1 . • If r\equiv 3\;\text{mod}\;6 , then a_{r}=b_{r}-1=c_{r}-1 . • If r\equiv 4\;\text{mod}\;6 , then c_{r}=a_{r}+1=b_{r}+1 . • If r\equiv 5\;\text{mod}\;6 , then b_{r}=a_{r}-1=c_{r}-1 . Proof. We prove this via induction on r : Basis Case \mathbf{0\leq r\leq 5} : Below is a table of values for a_{r},b_{r},c_{r} when 0\leq r\leq 5 : r a_{r} b_{r} c_{r} 0 1 0 0 1 1 1 0 2 1 2 1 3 2 3 3 4 5 5 6 5 11 10 11 Inductive Step: Assume r\equiv 0\;\text{mod}\;6 , the lemma is true for all r^{\prime}2 then a_{\alpha+d}+1=b_{\alpha+d} , Because b_{\alpha+d}\equiv b_{\alpha}\;\text{mod}\;m, and b_{\alpha}=a_{\alpha}-1 , this implies a_{\alpha+d}\not\equiv a_{\alpha}\;\text{mod}\;m , which again contradicts d=P_{m}(3) . Therefore 6|P_{m}(3) for m>2 . ∎ 4. Possible Periods for n=3 We can now prove Lemma 2 . Proof of Lemma 2 . Assume \mathbf{u}=(x_{1},x_{2},x_{3})\in\mathbb{Z}_{m}^{3} satisfies the sum condition. (1): It suffices to show if D(\mathbf{u})=(x_{1}^{\prime},x_{2}^{\prime},x_{3}^{\prime}) then x_{1}^{\prime}+x_{2}^{\prime}+x_{3}^{\prime}\equiv 0\;\text{mod}\;m . Notice x_{i}^{\prime}=x_{i}+x_{i+1} for i=1,2 and x_{3}^{\prime}=x_{3}+x_{1} . Therefore x_{1}^{\prime}+x_{2}^{\prime}+x_{3}^{\prime}=2(x_{1}+x_{2}+x_{3}), which is equivalent to 0\;\text{mod}\;m . (2): First, D^{2}(\mathbf{u})=(x_{1}+2x_{2}+x_{3},x_{2}+2x_{3}+x_{1},x_{3}+2x_{1}+x_{2}). Notice that x_{i}+2x_{i+1}+x_{i+2}=x_{i+1}+(x_{1}+x_{2}+x_{3}), which is equivalent to x_{i+1}\;\text{mod}\;m for 1\leq i\leq 3 , where the subscripts are reduced modulo 3 . Therefore D^{2}(\mathbf{u})=(x_{2},x_{3},x_{1}) or H(\mathbf{u}) . (3): Here, D^{6}(\mathbf{u})=H^{3}(\mathbf{u}) , which is \mathbf{u} . Therefore \mathbf{u}\in K(\mathbb{Z}_{m}^{3}) . (4)-(5): From the proof of (3), \text{Per}(\mathbf{u})|6 . Therefore, \text{Per}(\mathbf{u})=1,2,3\;\text{or}\;6 . Assume (x_{1},x_{2},x_{3}) does not satisfy the uniformity condition and is not \vec{\mathbf{0}} . If \text{Per}(\mathbf{u})=1,2 , then \mathbf{u}=D^{2}(\mathbf{u})=H(\mathbf{u}) implies x_{1}=x_{2}=x_{3} , which is a contradiction. If \text{Per}(\mathbf{u})=3 , then \mathbf{u}=D^{3}(\mathbf{u})=D(x_{2},x_{3},x_{1}) implies x_{2}+x_{3}\equiv x_{1}\;\text{mod}\;m x_{1}+x_{3}\equiv x_{2}\;\text{mod}\;m x_{1}+x_{2}\equiv x_{3}\;\text{mod}\;m. Plugging the third equation into the first yields 2x_{2}+x_{1}\equiv x_{1}\;\text{mod}\;m, or that 2x_{2}\equiv 0\;\text{mod}\;m. Similarly, 2x_{3}\equiv 0\;\text{mod}\;m and 2x_{1}\equiv 0\;\text{mod}\;m . If m is odd, then this only happens when \mathbf{u}=\vec{\mathbf{0}} , which contradicts our assumptions. If m is even, this can only happen when x_{1},x_{2},x_{3}\in\{0,\frac{m}{2}\} , in which case, \mathbf{u} is in the transition graph given in Figure 2 . (0,\frac{m}{2},\frac{m}{2}) (\frac{m}{2},0,\frac{m}{2}) (\frac{m}{2},\frac{m}{2},0) (0,0,\frac{m}{2}) (\frac{m}{2},0,0) (0,\frac{m}{2},0) Figure 2 . Part of Transition Graph for \mathbb{Z}_{m}^{3} , m Even Note that there might be other tuples in \mathbb{Z}_{m}^{3} whose Ducci sequences lead into this transition graph. Regardless, it is clear that if x_{1},x_{2},x_{3}\in\{0,\frac{m}{2}\} , then \text{Per}(\mathbf{u})=3 . ∎ This also yields the following corollary: Corollary 10 . For m>2 and n=3 , a Ducci sequence with cycle of length 6 always exists. We now prove Theorem 3 : Proof of Theorem 3 . Let n=3 and m be prime. We want to show that if \mathbf{u} does not satisfy the sum or uniformity condition, then \text{Per}(\mathbf{u})=P_{m}(3) . We first address the case where m=3 . Here, P_{3}(3)=6 . If you plot out the Ducci sequences for all 27 tuples in \mathbb{Z}_{3}^{3} , you would find that the only times when \text{Per}(\mathbf{u})\neq 6 for \mathbf{u}=(x_{1},x_{2},x_{3})\in\mathbb{Z}_{3}^{3} , is when x_{1}=x_{2}=x_{3} . Here, tuples meeting the uniformity condition have period 2 . Assume that m>3 is prime. Recall that 6|P_{m}(3) by Theorem 9 . Then a_{P_{m}(3)}\equiv 1\;\text{mod}\;m and b_{P_{m}(3)}=c_{P_{m}(3)}\equiv 0\;\text{mod}\;m . The first time that a_{r},b_{r},c_{r} satisfy this property is when r=P_{m}(3) , as if there was a smaller number where this happened s3 is prime, this yields b_{d}\equiv 0\;\text{mod}\;m. Since d\equiv 0\;\text{mod}\;6 , c_{d}\equiv 0\;\text{mod}\;m and a_{d}\equiv 1\;\text{mod}\;m . This contradicts P_{m}(3)>d . Therefore, \text{Per}(\mathbf{u})=P_{m}(3) . Suppose d\not\equiv 0\;\text{mod}\;6 . We have that d|P_{m}(3) , so take d_{1}=2d,3d,\;\text{or}\;6d so that d_{1}\equiv 0\;\text{mod}\;6 and d_{1} is as small as possible. If d_{1}6 by Lemma 2 . However, this is not always the case. To further investigate this, we present a few cases where n,m prime. Figure 3 provides cases where n>3 , n,m are both prime, and all tuples that are not \vec{\mathbf{0}} and do not satisfy the uniformity condition generate the maximum period. The value of P_{m}(n) and the period of all tuples that do satisfy the uniformity condition has also been included. n m P_{m}(n) x_{i}\neq 0 all equal 5 3 40 2 5 20 4 7 7 21 3 11 3 242 2 11 110 10 13 3 26 2 13 156 12 17 63856 8 29 24388 28 17 3 27880 2 17 136 8 19 3 373958 2 19 342 18 Figure 3 . n,m prime, three possible periods Figure 4 provides cases where n>3 and n,m prime and the only tuples with periods smaller than the maximum are \vec{\mathbf{0}} and those that satisfy the uniformity or sum condition. We are including the periods that the tuples satisfying the sum condition generate, in addition to the information provided in Figure 3 . For the cases in Figure 4 , note that the periods of the tuples satisfying the sum condition vary, even for a fixed n , unlike what we see when n=3 . n m P_{m}(n) x_{i}\neq 0 all equal \displaystyle{\sum_{i=1}^{n}x_{i}\equiv 0\;\text{mod}\;m} 5 7 240 3 80 13 420 12 140 17 360 8 180 23 2640 11 240 29 140 28 70 7 3 182 2 91 5 868 4 217 17 17192 8 2149 19 16002 18 889 23 6083 11 553 11 5 3124 4 1562 7 184866 3 16806 13 4084212 12 680702 17 7809208 8 1952302 19 27237078 18 3026342 13 7 509808 3 169936 11 7676760 10 1535352 Figure 4 . n,m Prime, four possible periods We now discuss how we made these conclusions. First, we run a program in MATLAB (from [ 16 ] ) that can determine the value of P_{m}(n) for a given n and m . Observing Figures 3 and 4 , the value of P_{m}(n) typically increases as n,m increase. As a result, for many n,m , P_{m}(n) becomes too large for MATLAB to find, preventing us from testing these cases. Our current program is currently set to stop trying to calculate P_{m}(n) after the program determines it is larger than 15 million. Once we find P_{m}(n) , Lemma 1 of [ 1 ] tells us that \text{Per}(\mathbf{u})|P_{m}(n) for every \mathbf{u}\in\mathbb{Z}_{m}^{n} . We can then create a list of possible periods for \mathbf{u} out of the divisors of P_{m}(n) . We now consider the example of n=5 and m=7 to demonstrate our strategy. Because P_{7}(5)=240 , the possible periods are \{1,2,3,4,5,6,8,10,12,15,16,20,24,30,40,48,60,80,120,240\}. We also take advantage of Theorem 7 of [ 13 ] , which says L_{m}(n)=0 . This is notable because for any tuple \mathbf{u}\in\mathbb{Z}_{7}^{5} , we have D^{\text{Per}(\mathbf{u})}(\mathbf{u})=\mathbf{u}. If we wish to test whether a divisor of P_{m}(n) , d , satisfies \text{Per}(\mathbf{u})=d for some \mathbf{u}\in\mathbb{Z}_{7}^{5} , we want to determine if there exists (x_{1},x_{2},x_{3},x_{4},x_{5}) such that D^{d}(x_{1},x_{2},x_{3},x_{4},x_{5})=(x_{1},x_{2},x_{3},x_{4},x_{5}). We will start by testing 40 . Notice the values of a_{40,s} are as follows: s 1 2 3 4 5 a_{40,s} 1 2 2 2 2 If \text{Per}(x_{1},x_{2},x_{3},x_{4},x_{5})=40 , then (x_{1},x_{2},x_{3},x_{4},x_{5}) satisfies the system of equations a_{40,1}x_{i}+a_{40,2}x_{i+1}+\cdots+a_{40,5}x_{i-1}\equiv x_{i}\text{mod}\;7 for 1\leq i\leq 5 , or A_{1}\begin{bmatrix}x_{1}\\ x_{2}\\ x_{3}\\ x_{4}\\ x_{5}\par\end{bmatrix}=\begin{bmatrix}0\\ 0\\ 0\\ 0\\ 0\par\end{bmatrix} where A_{1}=\begin{bmatrix}0&2&2&2&2\\ 2&0&2&2&2\\ 2&2&0&2&2\\ 2&2&2&0&2\\ 2&2&2&2&0\end{bmatrix}. Since det(A_{1})=128 , A_{1} is invertible, so there is only one solution. Because D^{40}(\vec{\mathbf{0}})=\vec{\mathbf{0}} , this one solution must be \vec{\mathbf{0}} . Since \text{Per}(\vec{\mathbf{0}})=1 , we can determine that there is not a tuple (x_{1},x_{2},x_{3},x_{4},x_{5}) such that \text{Per}(x_{1},x_{2},x_{3},x_{4},x_{5})=40 However, we must also bear in mind that if there exists \mathbf{u}\in\mathbb{Z}_{m}^{n} such that \text{Per}(\mathbf{u})=d where d|d^{*} , and d^{*}|240 , then D^{d^{*}}(\mathbf{u})=\mathbf{u} . Therefore, this also confirmed that 2,4,5,8,10 are also not periods for any tuple in \mathbb{Z}_{7}^{5} because they are divisors of 40 . Now, let us test 120 and its divisors as possible periods. Here, we aim to solve the system of equations A_{2}\begin{bmatrix}x_{1}\\ x_{2}\\ x_{3}\\ x_{4}\\ x_{5}\end{bmatrix}=\begin{bmatrix}0\\ 0\\ 0\\ 0\\ 0\end{bmatrix} where A_{2}=\begin{bmatrix}4&6&6&6&6\\ 6&4&6&6&6\\ 6&6&4&6&6\\ 6&6&6&4&6\\ 6&6&6&6&4\par\end{bmatrix}. Because det(A_{2})=448 , A_{2} is not invertible so there are more solutions than \vec{\mathbf{0}} . However, recall that if \mathbf{u} satisfies the uniformity condition, then \text{Per}(\mathbf{u})=3 . As a result, if 3|d , then D^{d}(\mathbf{u})=\mathbf{u} . Since 3|120 , all tuples satisfying the uniformity condition are a solution to this system of equations. We now want to confirm that the only solutions to this system of equations are \vec{\mathbf{0}} and tuples satisfying the uniformity condition. To begin, we make the assumption that x_{1}=1 and rewrite our system of equations, with the goal now being to solve \begin{bmatrix}6&6&6&6\\ 4&6&6&6\\ 6&4&6&6\\ 6&6&4&6\par\end{bmatrix}\begin{bmatrix}x_{2}\\ x_{3}\\ x_{4}\\ x_{5}\end{bmatrix}=\begin{bmatrix}-4\\ -6\\ -6\\ -6\end{bmatrix}. Since det(B)=48 , B is invertible. So there is only one solution to the system of equations when x_{1}=1 , which is (1,1,...,1) . We can repeat this process when x_{1}=j for 2\leq j\leq 6 and find that x_{i}=j for all 1\leq i\leq 5 . In other words, only \vec{\mathbf{0}} and those tuples satisfying the uniformity condition, which we already know have period 3 , are solutions to this system of equations. Therefore, 120 is not a period for any tuple, as well as 2,4,5,6,8,10,12,15,20,24,30,40,60 . Using these methods, we can eliminate all of the periods except for 80 . The a_{r,s} coefficients in this case are s 1 2 3 4 5 a_{80,s} 3 2 2 2 2 When setting up the system of equations, each equation becomes 2(x_{1}+x_{2}+x_{3}+x_{4}+x_{5})\equiv 0\;\text{mod}\;7. Therefore, if x_{1}+x_{2}+x_{3}+x_{4}+x_{5}\equiv 0\;\text{mod}\;7 , then D^{80}(x_{1},x_{2},x_{3},x_{4},x_{5})=(x_{1},x_{2},x_{3},x_{4},x_{5}). Since we eliminated the other divisors of 80 as periods, \text{Per}(x_{1},x_{2},x_{3},x_{4},x_{5})=80 . Therefore, all tuples satisfying the sum condition have period 80 . These methods were used to create Figures 3 and 4 . Finally, we provide a few examples of n,m prime that fall into neither of the cases presented in Figures 3 or 4 , which we determined using the above methods. Namely, we found divisors of P_{m}(n) that had solutions to their system of equations that were not \vec{\mathbf{0}} , did not satisfy the uniformity condition, and did not satisfy the sum condition. Figure 5 provides these examples, as well as the periods that came up as exceptions. n m P_{m}(n) x_{i}\neq 0 all equal Exceptions 5 11 10 10 2,5 19 90 18 45 31 30 5 3, 15 7 13 84 12 28 11 23 22 11 11 13 5 312 4 156 23 158158 11 11297 23 3 177146 2 88573 Figure 5 . n,m prime, Exceptions Out of these exceptions, we are able to look more in depth to the cases when n=5 and m=11,19 and when n=7 and m=13 , as \mathbb{Z}_{m}^{n} is small enough for MATLAB to provide a file with all of the tuples in \mathbb{Z}_{m}^{n} and their periods in a reasonable amount of time. For all of these cases, all tuples that generate a smaller period satisfy the sum condition. However, there were tuples that satisfied the sum condition that generated the maximum period. We now look more into the case where n=5 , m=11 . Suppose that you take a tuple in \mathbb{Z}_{11}^{5} with period 5 and you allow S_{5} to act on that tuple such that for \phi\in S_{5} , \phi\circ(x_{1},x_{2},...,x_{5})=(x_{\phi(1)},x_{\phi(2)},...,x_{\phi(5)}). Let J be the set of all \phi\in S_{5} such that if \text{Per}(x_{1},x_{2},...,x_{5})=5 , then \text{Per}(x_{\phi(1)},x_{\phi(2)},...,x_{\phi(5)})=5. By using MATLAB to find J , we determined that J\cong D_{10} . This can similarly be shown for when n=5 , m=19 and n=7 , m=13 and the period that appeared as exceptions for each case: Let d be the smaller exception period for \mathbb{Z}_{m}^{n} where either n=5 and m=19 , or n=7 and m=13 . Let J to be the set of all \phi\in S_{n} such that if \text{Per}(x_{1},x_{2},...,x_{n})=d , then \text{Per}(x_{\phi(1)},x_{\phi(2)},...,x_{\phi(5)})=d . Then J\cong D_{2n} . This is not the case for the tuples in \mathbf{u}\in\mathbb{Z}_{11}^{5} such that \text{Per}(\mathbf{u})=2 . All of these tuples satisfy \mathbf{u}=z*(1,9,4,3,5) where z\in\mathbb{Z}_{m} , z\neq 0 . For the cases where n,m are prime that we have investigated, we found one case where n,m are prime and every tuple that satisfies the sum condition generates a smaller period than the maximum and that also has other possible periods. This is when n=7 and m=11 . Here, P_{7}(11)=1330 and tuples that satisfy the uniformity condition have period 10 . There are 1330 tuples with period 19 , all of which satisfy the sum condition. The rest of the tuples satisfying the sum condition have period 133 . There are 13300 tuples with period 190 , none of which satisfy the sum condition. Let J_{1} be the set of all \phi\in S_{7} such that if \text{Per}(x_{1},x_{2},...,x_{7})=19 then \text{Per}(x_{\phi(1)},x_{\phi(2)},...,x_{\phi(7)})=19 . Let J_{2} be the set of all \phi\in S_{7} such that if \text{Per}(x_{1},x_{2},...,x_{7})=190 then \text{Per}(x_{\phi(1)},x_{\phi(2)},...,x_{\phi(7)})=190 . Then we find that J_{1}=J_{2} and they are isomorphic to the Frobenius group of order 21. 6. Periods when n=m is Prime We draw our attention back to Figure 3 to note that all of the cases when n=m for n prime, 5\leq n\leq 19 fall into the case where all tuples except \vec{\mathbf{0}} and those that satisfy the uniformity condition generate the maximum period. This leads us to Theorem 4 . Before we prove it, we prove a lemma. Lemma 11 . Consider an j\times j matrix that follows this pattern: B=\begin{bmatrix}-1&1&-1&1&\cdots&\pm 1\\ 0&-1&1&-1&\cdots&\mp 1\\ 1&0&-1&1&\cdots&\pm 1\\ \vdots&&&&\ddots&\vdots\\ \mp 1&\pm 1&\mp 1&\pm 1&\cdots&-1\end{bmatrix}, where \pm is positive if j is even and \mp is positive if j is odd. Then det(B)=\begin{cases}1&j\;\text{is even}\\ -1&j\;\text{is odd}\end{cases}. Proof. We prove this via induction on j , with j=2 and j=3 as our basis cases. Inductive case: Assume that the hypothesis is true for j^{\prime}