Optimizing Logistics Route Planning And Monitoring System

Should be concise and descriptive. Be precise enough to suggest the nature and scope of the project and concise enough to be referred to quickly and easily. The title should also be “catchy” to attract the readers’ attention The title page should also include the units associated with your project. For example if you are doing a 48-credit point project you should state that this project involves IFN701 and IFN702.

The Importance of Logistics in Modern Times

The day to day habits of the people, like entertainment, travel have been changed with the development of time and advancement of the science and technology. Prominently changing habit these times, are shopping, a big transition from the traditional and physical shopping to online shopping. Eventually, there is an explosive growth to the online stores and distribution services have been an added chunk to large number of traditional shops. The key and essential role hence has been played by the logistics. The functionalities of logistics have been appended with warehousing, packaging, traffic, warehousing, with significance of more than providing shipping service to the customers. Hence, an optimized route plan not only reduces the cost of the companies, but also increases the clients’ satisfaction. The new methods have also need to be integrated with the real time events, for better monitoring system (Alshamrani & Bahattab, 2015). Hence, monitoring system is a crucial part that focuses not only on turning data into real pictorial graphics, but also emphasizes the safety vehicles, distribution of road conditions and real time tasks.

The context for the project is to study, analyze and compare the current logistics route plans and create a new route planning so that its efficiency of companies can be improved (Abraham et al, 2012). A new route planning is proposed after the existing problems existing and faced by the distribution firms.

The purpose of the ‘monitoring system in logistics planning and optimization’ explores the real time position of the logistics vehicles is detected and analyzed, like speed and mileage (Arz et al, 2013). The modern logistics requirements cannot be adapted by the traditional planning, like when a task is performed by the logistics vehicle and a new task or an accident happens on the road, the problem or issue, would be how to develop and propose a better alternate route, still justifying the cost (Zhang et al, 2017). It needs the analysis of the current road conditions on real time basis and plan should be developed towards a better cost-effective path.

So, after the literature review and analysis, the final deliverable will be a better and efficient monitoring system that is built the Application Interface developed by Google Maps. This new system should be able to receive all the real time information for the vehicle, such as fuel consumption, speed, loading capacity and any relevant information to Google Maps, along with the feedback. A new path will be created by refreshing the information regarding the vehicle information, and relative condition of the road to the customized destination from the customized start, for every 30 seconds. As a result, the monitoring system reflects the direction towards distribution vehicle’s driving direction, for stakeholders to check, on the Google Map (Antsfeld & Walsh, 2012).

The Purpose of the Study: Route Planning Optimization and Monitoring System

Monitor system in logistics planning and optimization involves basically a study of the route algorithms applied in the same companies and further a better route algorithm is to be proposed after improving, through developing a new series of applications. It should help and support courier corporations to improve the driving path design, after analyzing the routes, intuitively, for artificially made express cars.

Literature review is performed by searching and exploring the data about the monitoring system. A crucial part concerned and focused in this context is the algorithms that emphasize the distribution of road conditions, real-time tasks and safety of vehicle, so that data analysis helps to make decisions towards benefiting the society, by benefiting the companies and customers. To do so, initially, the way the past algorithm or route planning schedules and methods are studied, followed by the ways the methods to become better efficient methods, while addressing the real time events (Bader et al, 2011). Further better methods proposals are explored, to lower the shipping cost so that the short notice of customers, vehicle issues and route availability demands can be addressed.

The existing literature has given the clue about the three difficulties that the distribution corporations do face, usually, in the logistics in the cities. The first difficulty is the relationship between travel time and time of the day, which can be solved by exploiting a lot of data that is available today. The second problem addressed is dynamics in cities. The third problem is the limitations of the infrastructure, like few parking lots, narrow streets and one way.

Existing literature shows the current route planning of the current logistics and also helps to investigate new ways and methods of planning route so that the efficiency of these companies can be improved. After improvement of the existing route planning, a monitor system is proposed to monitor the information of the traffic, with reference to the speed routes, mileage, vehicles of the company vehicles.

A new logistics industry generation development resulted in distribution vehicles to increase in the same industry, gradually, however, gradually, the same also has brought increased pollution, automobile exhaust and index of air quality is gradually increased. Modern industry of logistics stands as a significant China’s economic development component, so that the common economy and environment sustainable development can be achieved, by necessitating reducing the consumption of energy, in order to increase the green area (Bast et al, 2007). The energy consumption is basically and mainly in the process of distribution, in the logistics industry, towards achieving low energy consumption, the distribution route problem is resolved and improved, using effective and reasonable algorithm, to realize the optimization of distribution path to be a significant sustainable development foundation and stands as the ultimate goal.

The Challenges of Logistics Route Planning and Optimization

And another dimension for the logistics evolution challenges are various methods used to predict the system of freight transportation, while representing and analyzing. The challenge is found in a more detailed multimodal transportation system network representation and the more refined demand development, assignment models and mode choice (Bast & Storandt, 2014). This goal can be achieved by focusing on these methodologies key component, the representation and shortest paths computation. Hence, the route algorithms can be made user optimal with the best route algorithm for shortest path routine. The best and efficient algorithm should be explored appropriately, for appropriate applications, in the logistics industry.

The basis for finding the shortest path for travelling, in the logistic industry is the shortest path problem. An important issue to consider is the shortest paths computation on networks, in this distribution that represent movements of freight through multimodal systems (Batz et al, 2013). The paths modeling should take only the modes to take into account, for making up the chain of transportation, but also the way the networks of corresponding career and the operation of specific logic distribution systems (Bast et al, 2007). A shortest path algorithm with monitoring system has to be proposed and developed after studying and analyzing the existing shortest path route algorithms.

The usual and regular problems considered are movement of, product flows that are potentially different, in between the given points of origin and destination, over networks of multimodal transportation. The transportation supply representation takes the multimodal network form (Barrett et al, 2008). In this context, mode is transportation means with certain characteristics like, vehicle capacity, type and measures of cost.

A shortest path is needed to visit a certain intermediate nodes set, in a particular sequence, like particular vehicle’s itinerary. A conventional algorithm of shortest path is applied to find shortest path, from the origin to each of the intermediate node, followed by intermediate paths till the point of destination (Bast et al, 2009). The problem needs to consider the path to use one successive node sets that are not contiguous necessarily and permit the modes set’s any usage, between such intermediate nodes successive set.

It should be noted that the intermediate node sets, in most of the cases as do not form disconnect of the network. However, it implies that the problem cannot be decomposed along sets of intermediate node. So, sub-paths have to be explored, till satisfying paths that satisfy all the found constraints. However, it is determined may be only in certain instances at the destination, in case a mode allowed in the first set of constraint provides a complete path, while constraints are not satisfied (Crainic et al, 2007).

The Proposed Solution: Real-time Monitoring System Using Google Maps

There are certain optimal algorithms existing for routing planning, such as IP formulation. However, it is infeasible, computationally for obtaining the routing planning solutions for the logistics applications. It is also indicated that generating optimal solution for larger size routing planning is almost impossible, within reasonable and small amount of time. Such problems can be solved, however, with heuristics, replacing the optimal algorithms (Bast et al, 2013). Near-optimal solutions can be obtained with several heuristic algorithms. Some of the important algorithms are genetic algorithms, greedy, 3-opt, 2-opt, neural network, simulated annealing though they have varied efficiencies from size to size and case to case. Such algorithms can further be improved through non-crossing method (Hao & Suling2009).

The key point and concept of the construction algorithm is to construct the shortest path by including the points in the tour, until the development of the complete tour, by including the points one by one (Batz and Sanders, 2012).

However, two or more points can be transposed in the initial tour, if possible, in order to give an initial solution improvement, in improvement algorithms. When improvement algorithm is considered, there are two stages possible. There can be two alternate considerations to choose. The first one is to choose one exchange, after considering all the possible exchanges, to obtain greatest savings and let the process to continue till best possible reduction, further (Bast et al, 2014). The second one is to make the exchange, as soon as a saving is made available and examine other exchanges possible and let the process to continue till the solution cannot be improved any further. The examples for the first consideration are 2-opt algorithms and 3-opt algorithms and the examples for the second considerations are greedy 2-opt algorithm and greedy 3-opt algorithm.

Hybrid algorithms make use of both construction algorithm and also improvement algorithms. It makes use of construction algorithm to gain an initial solution and further improvement of the algorithm is done with the help of improvement algorithm (Babenko, 2013). The best examples for these algorithms are neural network and simulated annealing algorithms.

Greedy algorithm is an improvement algorithm, in a simplest form. The process is initialized with departure node 1. Then all the distances to rest of the n-1 nodes are calculated. Then the next closest node is taken (Aifadopoulou et al, 2007). Current node is then considered as a departing node and nearest node is selected from the rest of n-2 nodes. The same process is continued till each and every node is visited, but only for once and then it goes back to the Node 1. Finally, the sequence is obtained as the best route, by the end or termination of the algorithm. Then, it is considered as the best solution (Abraham et al, 2013). This algorithm is easy to understand and implement as it is simple. And it helps to obtain the best solution, sometimes, where the size of the problem is small. In addition to that, it saves a lot of computational time, as there are no changes of nodes entailed.

The Shortest Path Problem in Logistics

2-Opt Algorithm

When the logistics routing planning is considered, for n routes or nodes, the 2-opt algorithm proceeds in three steps. In this context, greedy algorithm is used for setting objective function value, along with initial solution (Bast, 2009).  
Let z = Objective Function Value, S = Initial Solution, given by the user

1. Set S* = s and z* = z, where the increments i = 1 and j = i+1.
2. Let S` = Solution after exchange results and it has OFV z` < z*, and then set S* = S` and z* = z`. This step 2 is repeated if the condition j<n is satisfied. If this condition is not satisfied, increment the values of increments as i=i+1 and j = i+1. If the increment value of i is still lesser than n, iterate with the step 2. If not jump to step 3.
3. If value of S is not equal to S*, then set z=z* and S=S*. Then initialize i with 1, as i=1 and j=i+1. Then get back to the step 2. If not, the best solution stands as output S*. Then end or terminate the process.

When the algorithm is observed, it only attempts pairwise exchange. Here, Nodes 1 and 2 transposition is considered, initially. If the OFV of the resulting solution is smaller compared to that of the initial solution, then for future, it is stored as a candidate. Otherwise, it considers Nodes 1 and 3 transposing, after discarding the previous step (Crainic et al, 2009). If a better solution is generated from this exchange, then for future consideration, it is taken as a candidate. Otherwise it is discarded. Hence, the algorithm discards the best and previous solution, when it finds a better solution. The same process is continued till consideration of all the pairwise exchanges.

When there are total n nodes, since each of the node is exchanged with other n-1 node, the total different exchanges would be n(n-1)/2, which are conducted in the step 2. The most improvement in the OFV is the final solution retained, by the end of the iterated step 2. Considering it as a new solution, the step 2 is repeated for finding other better solution (Cohen et al, 2003). Going forward, there is no improvement found in the current best solution and the end of the algorithm occurs. The rest of the solution is found to be the best one to the user.

Greedy 2-Opt Algorithm

The algorithm, Greedy 2-opt is a bit variant of the algorithm, 2-opt. This algorithm is also processed in three steps.

Let z = OFR (Objective Function Value) and S = Initial Solution.

1. Set z* = z and S*=s and set the initialize i=1 and j=i+1.
2. Perform transpose of the nodes, Node j and Node i, in case i<j. The resulting z* value is to be compared with OFV z. If z is greater or equal to z* and still j is lesser than n, then repeat the same step 2, after incrementing j = j+1. Otherwise shift control to step 3.
3. If z<z*, then set z*=z, S*=S and continue for i=1 and j=i+1, then shift control to step 2. Then increment j=j+1 and i=i+1, if j=n and z is greater than or equal to z*. Otherwise, set the S* output as the best solution and then process will be terminated.

Analysis

Greedy 2-opt algorithm, like 2-opt algorithm performs the pairwise exchanges. Transposing of Node 1 and Node 2 is considered, initially, if the OFV resulted is lesser compared to the previous result, the transpose of these two nodes is performed immediately. Otherwise, the algorithm considers Node 2 and exchange is then evaluated and then the process is continued till the improvement is found. The algorithm considers transpose of Node 1 and Node 2, as an initial solution (Abraham et al, 2011). This algorithm is continued until any further improvement of the solution is impossible. The algorithm considers making permanent exchange, after finding the improvement and so comparatively 2-opt algorithm, this algorithm needs lesser computational time. However, the algorithm obtains the solution that is a bit worse compared to the 2-opt algorithm.  

Optimal Algorithms for Routing Planning

3-Opt Algorithm

The only difference or deviation between the 2-opt algorithm and 3-opt algorithm is that this algorithm performs two nodes transpose simultaneously, through both are similar in other aspects (Batz et al, 2012). Transposing of the nodes is possible in two different ways, j-k>->i->j and i->j->k-i.

When a transpose is considered, the following three steps are executed.

1. Set z*=z, S* = S and set increments i as i=1, j=i+1 and k=j+1.
2. When transpose of Node i, Node j and Node k is considered, in the fashion of ‘i->j->k->i’, set S*=S` and z*=z`, if the S`, a resulting solution has smaller value of z`<z*. Attempt to set k=k+1, if k<n, otherwise, set j=j+1.then set k=j+1, if j<n-1. Otherwise set k=j+1, j=j+1 and i=i+1. Attempt to repeat the step 2, if i<n-2. Otherwise proceed with the step 3.
3. Set z=z*, S=S*, i=1, j=j+1 and k=j+1, if S is not equal to S* and then proceed to the step 2. The output S*, otherwise is the best solution and the entire process is terminated  (Bauer, 2012).

Greedy 3-Opt Algorithm

Greedy 3-opt algorithm is similar to that of the greedy 2-algorithm, except that it makes the exchange of 3-node permanent, while the OFV resultant is better compared to the current OFV and a new transposition is considered as an initial solution and the algorithm is repeated (Abraham et al, 2010). When a transpose of ‘i->j->k->i’ is considered, the process is executed in the following steps.

Let z = OFR and S = initial solution

1. Set z*=z and S*=S, i=1, j=i+1 and k=j+1.
2. When transpose of the Node i, Node j and Node k is considered, in the fashion of ‘i->j->k->i’, if S`, a resulting solution is smaller z`<z*, then set S*=S` and z*=z. Then set k=k+1, if k<n. Otherwise, perform setting of j=j+1. Then set k=j+1, if j<n-1 and otherwise set i=i+1, j=j+1 and k=j+1. Repeat the step 2, if i<n-1, otherwise jump to the step 3.
3. Set z*=z and S*=S, i=1, j=i+1 and k=j+1, then jump to the step 2. Set i=i+1 and j=j+1, if z is greater than or equal to z* and k=n, then jump to the step 2. Set i=i+1 and jump to the step 2, if j=n-1 and z is greater than or equal to z*. The output S* is otherwise becomes the best solution and the process is terminated.

When the graphs of the graph theory and geographical network, graph theory is widely used and applied in logistics. As a result shortest path algorithm, in association with graph theory is applied to solute vehicle routing, as a based effective algorithm (Xiao-nian & Xiao-liang, 2009). When logistics route is considered, it is like a simple edge related to the graph. And here, vertex is considered from the series of transporting or loading locations. The distance between the two locations, traffic expenses and journey time are considered as a weight of edge (Yang & Jian-ya, 1999). Though the algorithm of the shortest path is an antiquated and traditional method, it still stands as a hot issue, in terms of research of optimal path. There are around 17 different algorithms proposed for graph theory based shortest path and only three types of algorithms, TQQ, DKD and DKA tested by the experts. Here, DKD and DKA algorithms are based on Dijkstra algorithm. Both of these algorithms are almost the same, with only one difference that the way they are implemented.

The key concept of the project is to explore and understand the route planning algorithm. So, when the starting point and destination points, on the road are given with no barriers existing on each location, distribution lines optimal path has to be obtained (Nabil et al, 2010). Here, the primary concern that is encountered by the distribution center is the way the optimal routes are sought and the core algorithm, of which would be the algorithm for the shortest path. Let us establish a model for geographic network, suiting for the routes planning of logistics. Among the three experts tested shortest path algorithms, DKA, DKD and TQQ, and the former two are based on Dijkstra algorithm, while the later one is TQQ is graph growth theory(Hao & Suling, 2009).

The Construction Algorithm for Shortest Path

Dijkstra algorithm is applied to solve the shortest path problem of single source, for a graph having the costs for nonnegative edge path that produces the tree for shortest path, as it is a graph search algorithm (Bauer et al, 2013). So, this algorithm is one of the best suitable for route planning. So, most of the current systems make use of this algorithms, to solve the shortest path issue, as a basic theory, though there are varied systems for realizing Dijkstra algorithm, by varied systems (Hui et al, 2009). Usually, Dijkstra algorithm has the applications for minimum cost path computing, to all nodes, from the source node.

Dijkstra algorithm consists of the basic idea of exploring the shortest path gradually, to outside from the source point. Initially, each point is considered and assigned a number for each point. The numbers here express the shortest path weight from s (source point) to P (label, p) or shortest path weight upper bound to the point T (label T), from s. Each step is considered and T label is modified and the point with T label is altered to point having P label. Then the number of vertex having P label increases one, in graph G, then shortest path can be obtained to each point from s, in just n-1 steps (Delling et al, 2011).

The algorithm is further optimized, by expressing Dijkstra algorithm in another way. For example, if each point consists of label pair, (dj, pj), where pn is the j’s front point and dj is the shortest path length, measured from the starting point, in the shortest path to j from s (Bast et al, 2010). The simple solving process for the shortest path algorithm to point j from the starting point s, is as the following (Akiba, 2014).

1. Initialization – Here, the starting points are set as,
   1. Starting point is to be set as, ps = null and ds = 0
   2. All other points to set as di = ∞ and what is pi value?
   3. Starting point s is to be marked as k=s and unlabel all other points
2. Examine the distance in between the k, the marked point and j, the unlabeled point, which is connected directly to k. When lkj is considered as the distance of direct connection, in between k and j, then set dj = min[dj, dk+lkj].
3. Select the next point. Select the smallest i, from all the unlabeled points, in di, if di = [dj, j], and i is set as marked and selected as one shortest path point.
4. The front part of the i has to be found. Explore and find j that is directly connected from marked points to i, set it as pi=j and also as front point (Brandes et al, 2001).
5. Mark i. The algorithm is finished, when all the points complete marking or if the marking of target point is done, otherwise get back to the step 2, after setting k=i and continue.

According to the Dijkstra algorithm discussed above, the core step is to select an arc, from unlabeled points, with the shortest weight, to achieve the algorithm (Bast et al, 2013). It is a process of cyclic comparison. In case storage of the unlabeled points is done in an array or linked list in unordered form, all the points have to be scanned so that an arch can be chosen with the shortest weight. It will in turn affect the speed of computing, when the data is in larger amounts.

Heat sort, when combine with actual situation, is used to order the unlabeled points so that the shortest path and efficiency of the node are improved as the key heat sort process. The reason is as follows.

1. Get complete usage of heap data existing and reduce the data comparison frequency, greatly. Usually, the heat sort run-time is primarily consumed in initial heap construction. In this algorithm, only one heap is to be constructed in the entire seeking process to find the shortest path, with the heap sort of n times. So, the drawback in the heap sort can easily be overcome this way and the advantage is clearly shown.
2. Shortest path value changes in the Dijkstra algorithm are usually, smaller compared to the original value. Hence, small root heap is used and is also called as heap root. Here, this root is the node with the value of smallest keyword (Delling et al, 2013). After modifying some node’s keyword, while heap operation refreshment, it is to be determined whether there is need for the node, for adjustment of the position to the parent node of it (Bauer et al, 2013). Hence, the heap rearrange operations in this algorithm are better efficient and more simple compared to the traditional heap sort.
3. When heap sort is considered, it requires only a record’s auxiliary space and for merge, it requires O(n) and quick sort, it needs O(log n).

Dijkstra algorithm can further be optimized by optimizing the distribution lines path with barriers. An example can be shown and discussed with the known conditions (Akiba  et al, 2013). However the degree of optimality does change with the varied conditions. Optimal path can be obtained with certain method, in changing conditions.

Let us suppose that the goods need to be delivered to the destination D from the initial location, O, for a logistics distribution center. Logistics distribution center is regarded as the urban traffic network, ideally and shown as a plane graph so that shortest path in between O and D can be calculated, so that costs and time of transport can be saved (Cattaruzza et al, 2017). However, the path is not smooth enough in many instances, and it may be a problem then. When a truck of transportation goes in shortest path and reaches the location A, and the driver finds it impossible to pass through, because of the breakdown of the road, in the location B, which is ahead, the route must be changed and has to ensure that the path selected is optimal.

Mathematical Model

Let us suppose that G is a plane graph marked for an urban traffic network. In graph G, each of the location in the route of traffic corresponds to each of the vertex (Brandes et al, 2005).

So, set G = (V,E) having weighted edge as a undirected graph.

Here,

E = Set of Edges

V = Set of Vertices

And in a triangle, the sum of the two edges length is greater than the third edge. So, this concept is known as triangular inequality. Here, D is the destination and O as the starting point, for transport and SP is the shortest path having no barriers (Alluri  et al, 2014). So, E(SP) shows the costs spent for the transport on the shortest path. To conduct the discussion, there are two assumptions made as the basis.

1. G graph is connected still, after choke point removal.
2. And the latter point cannot get through, caused from the blockage, as transport vehicle reaches the previous point.

Design of Algorithm

Data structure having two sets for graph vertices to deposit is used for Dijkstra algorithm. The s set shows the marked nodes set and the (G-S) set shows the unlabeled nodes set (D-Angelo et al, 2012). The node label shows the shortest distance, in between the source point and this specific point.

The algorithm has a major concept is to select the smallest label node W from the G-S set then place it in the S set. Then adjust all the node v path value (T[v]), passing through the node W and gets connected to the set G-S contained source node. In case the node v path value is greater than the sum of labeled node W path value and distance in between W and v, then all nodes path value is to be adjusted that are connecting to the v node (Akiba  et al, 2013). This process is to be repeated until every point is entered into the S set.

The discussion above can be analyzed in the way that after operating Dijkstra algorithm, each and every node is labelled. Shortest distance is represented with WOA in between the nodes A and O and W(SPAD) represents the distance in between node D and A, along the SPAD optimal path, Wij represent weight of edge, to node j from node i. Here, T[v] shows the node v label value (Brunel et al, 2010).

It shows the algorithm as follows.

1. Use Dijkstra algorithm, for the plane graph G given, for getting the SP optimal path to the end D from the source node O.
2. Compute a sub-graph G-{B}, when the vehicle reaches the node A and it gets impossible to pass through the node B, as the barriers are present.
3. Dijkstra algorithm is to be reused to get the SPAD optimal path, to the node D from node A.
4. The cost to node D from node O is the sum of cost W(SPAD) to node D from node A and WOA (to node A from node O).

Practical Implementation

Let us consider a practical situation for a logistics company transport. The company needs to transport a fresh fruit batch to the city B, from the city A. The fruit needs to be delivered to the destination, in optimal path to reach in the shortest time, to maintain the value and freshness of the fruits. The task is to obtain the efficient route SP with an efficient route planning, to complete within the W(SP) cost.

Let the staff to start from the A city and move through the SP route and obstructed in the city V, because of the floods caused road congestion (Bauer et al, 2010). Here, there are only two choices for the staff. First choice is staying put for the repair of the bridge and the second choice is to find an alternate way to move on to the destination city, B. However, it is not easy as it is complex and comprehensive problem. At times, staff prefers to walk through the way of optimal path and so staying put would become better and appropriate, according to the calculated optimal path, in advance, so that there are no other expenses incurred by the distribution team, towards choosing another way and path (Barett  et al, 2009). However, it has an adverse factor, in terms of consumption of time and cost. The fresh fruits value is the feature of fresh, so the fruit must transport at the shortest time, to the destination. But, the fruits’ value of freshness would be lost if stays put. So, an alternate path has to be selected to move on, by the staff. And now the task is to reselect the path, for ensuring arrival of time. The graph for road is shown as follows?

Considering this context, let us assume that the cost of the transport to finish from the start is proportional to the time to finish from start. So, to reach the destination city B, from the starting city, V, the optimal path can be SPAVB, based on the algorithm discussed above (Brodal & Jacob, 2004). Here, Dijkstra algorithm is used to obtain the SPVB optimal path to reach the city B, from the city V. And the shortest path and time is obtained as WAV + W(SPVB).

For a diversified transportation era, cost cannot be controlled by the logistics company, if just distance between the source and destination cities is considered (Xiao-Yan & Yan-Li, N.Y). So, staff and the logistics company, in the distribution process, need to consider not just the route, however, there are also other several factors, like weather, road condition, holiday and others.

Genetic algorithm simulates the genetic solution process and natural elimination process in biologic solution, as it is a computational model. The usual genetic algorithm computation is an iterative process, simulating the natural elimination and genetic selection process, in biologic evolution (Botea & Harabor, 2013). Candidate solution is retained and gets ranked, for iteration and the ranking is done on the basis of the quality. Unqualified solution is screened out through fitness value. New candidate solutions are then estimated for the next generation, by performing the genetic operators, like mutation, crossover, inversion and translocation on such qualified solutions. The same process is iterated till it meets certain convergent condition. The final general search method would be the genetic algorithm. The algorithm makes use of the genetic operator’s analogs, on states’ population to find such states that have greater fitness values, in a search space (Goldberg, 1989)

Genetic algorithms are used to explore the available range of options. Its ability to search a space for solution and focus on the promising criteria combinations, allow them to be suitable to several problems associated with the spatial decision, ideally (Bohmova et al, 2013). Search problems are such that the routing problems in the navigation systems of car, to find the optimal route to the destination from an origin, within a limit of time, on a road map.

When optimization problems are considered methods of genetic coding and genetic operators are designed, in advance to satisfy the constraints of individuals. However, the CSP (Constraint Satisfaction Problems) objective is to ensure that the individual has satisfied constraints, to be the fitness value. Crossover and mutation conducts the usual search in genetic algorithm, based on theory called neo-Darwinian evolution. Generally, crossover is considered as a robust means of search. The partial solutions may be inherited by offspring, with no conflict from the parents, but does not inherit any information to decide the partial solutions containing genes, in terms of availability. From a point of view of search strategy, it refers that selection of variables is done randomly and then assigning values are also selected randomly, from genes that contain in the population. So, in genetic algorithm search by crossover is not considered to be efficient (Delling et al, 2008). And moreover, mutation is considered as itself as a random search method. However, the most important characteristic of the genetic algorithm is efficiency of global search.

Strategies

Let us consider N = Node set

C = Cost set

E = Edge set

A(directed) graph G = (N,E,C)

Here, E = subset of the N*N cross product.

In E, each element (u,v) is an edge joining to node v with node u.

Each of the element of edge (u,v) is related with a cost C(u,v).

Cost C(u,v) obtains values from the real numbers set. Both node u and node v are neighbour values, if edge (u,v) is within E (Bielli et al, 2006). Node degree is the neighbouring node’s number. In a graph, the path to a destination node d from a source s, is a nodes’ sequence, (v0,v1,v2,.... vk) and here, d = vk and s = v0 and the existing edges (v0, v1), (v1, v2),.... (vk1, d) in E. The total number of edges in a path is known to be path’s cardinality. So, path’s cost is taken as sum of the edges’ cost, which is,

For node u to node v, the optimal path is the path having a smallest possible cost. The path set to node v from node u is represented by the (u, v) paths. A path suffix is gained by nodes and edges removal from the path’s beginning.

For instance, when a path through (v1, v2.... vk) is considered to be a path suffix, through (v0, v1, v2,... vk). S, which is a shortest_path_tree is a shortest path s collection to all nodes, from s, the source node, in the graph, having s as the root node (Dehne et al, 2012). The graph diameter is considered as the shortest path’s largest cardinality, in between any of the pair nodes (Bauer et al, 2012). In case any path exists to all other nodes, from all nodes, in the graph, it should be considered that the graph is connected.

Let us focus on the connection of the graphs. A cost estimator of path is considered as a function f (u, v), in a graph, which computes an optimal path estimated cost, in between the node v and node u. If the optimal path exact cost is computed, a path cost estimator becomes a perfect estimator, between specified two nodes (Berger et al, 2009). A cost estimator of path is admissible in case it always under the path’s cost estimates, like f (u, v) is lesser than or equal to path (u, v), for all the paths present paths (u, v). In case f (suffix (path)) is lesser than f(path), then the cost estimator of the path is known to be monotonic, for all the suffixes and paths. Path problems of single pair are divided into any path problems and shortest path problems (Bauer et al, 2012). Definition of shortest path problems can be understood as the following. When a graph, G = (N, E) is given, and node v and node u in N, then in order to find the shortest path in between the nodes v and u, it will be the path having smallest possible cost. Any problems related to the path are defined by shortest path constraint removal from the problem of shortest path. 

When a practical system is considered during driving and when congestion of traffic changes, then the re-evaluation of route should be done, before car reaches the following intersection (Golden, 1976). These search problems can have representative solutions as Dijkstra algorithm.

Greedy algorithm has an advantage that it is simple to understand and so to implement. This algorithm helps to obtain the best solution, usually, when the size of the problem is very small, because of the fact that it does not entertain any of exchange of nodes and so saves a lot of computational time (Bauer et al, 2010). However, it suffers with many issues and efficiency of the final solution, especially, when the size of the problem is very big. However, improvement of this algorithm can be done, when it is enhanced with the other algorithms, like 2-opt and 3-opt.

Since Dijkstra algorithm is considered to be an exact and suitable algorithm, as it is able to determine an optimal route always. However, it cannot guarantee that it can meet realistic deadline.

However, when genetic algorithm is considered, it always shows the solutions, during the search, in a population and it can use other solutions and provide alternative routes, in the shortest time (Berger et al, 2010). Hence, genetic algorithm can be used to find quasi-shortest time and easiest to drive route so that destination can be reached within a stipulated time. So, the algorithm helps to produce the candidate routes and choose as well so that meeting of deadline can be guaranteed and so can satisfy the constraints, in terms of ease of driving.

Genetic algorithm demonstrates better and favourable performance, for global optimization as a high efficient search strategy, on solving the problems associated with combinatorial optimization. Genetic algorithm, when comparing to the other traditional search algorithms, is able to accumulate and acquire the knowledge needed about the search space in the process and control the entire process of search with self-adaption, through the technique of random optimization. So, most probably this algorithm have applications to provide the global optimal solution having no combination explosion trouble encountering, resulted from disregarding the inherent knowledge present in the search space (Bauer & Delling, 2009). So, it is more like in solving the problems of combinatorial optimization and non-linear problems having non-differentiable objective functions and complicated constraints. It also necessitates the genetic algorithm application into the algorithms for GIS route finding.

However, the most important characteristic of the genetic algorithm is efficiency of global search. Genetic algorithm has a lower rate of search, it stresses random search, instead of the directional search. And still genetic algorithm helps providing favourable performance on the combinatorial optimization problems solution, as a search strategy, with higher efficiency, for global optimization (Raichaudhuri & Jain, 2010). Hence, the best problem of route selection, in the analysis of network is solved having genetic algorithm with fitness function selection, efficient encoding and different genetic operations. Final solution can be obtained from the crossover that can be the most significant operation. So, a better method for implementation can be designed effectively, in terms of complexity and time of computation. Additional efforts are done, on two different items, like enhancement of the algorithm’s adaptability, under dynamic constraints and expansion of the algorithm applications, into broader topics of GIS.

Google Maps API in simple terms is highly sophisticated algorithm, released by Google to integrate the services of the Google Maps into the websites of the users and it is not completely revealed by Google to the external world. So, the website can also overlay with certain site specific data. The application is built to retrieve the static images obtained from Google Maps and web services to perform geocoding, obtain elevation profiles and generate driving directions (Berger et al, 2011). It is so far a heavily used API web application development, used by more than 1,000,000 websites.

The Application Interface of the Google Map helps to integrate various services of Google Maps, such as Google Direction, Google Geocoding and Google Static Maps so that the integrated services can help to obtain various real time data, such as traffic, distance and other data.

Logistics vehicle system has got its monitoring system, best possible after customizing the Google Maps application for the same. This application is developed on the name of Google Biking Directions, on March, 2010, after adding the possibility to search for the direction for the vehicles through Google Maps. This algorithm and application can best be customized and applied for the logistic vehicles to find the shortest path to reach the destination, after exploring multiple and available paths and choosing the best, cost-effective and fast reaching path (Baum et al, 2014). The API calculates the elevation change, bike lanes, traffic, bike paths, preferred roads and optimal routes for vehicles to move on (Dehne et al, 2012). Though it was initially developed for biking, the algorithm also works for logistics vehicle monitoring system.

Considering the logistics situation, where the shortest path has to be chosen among the multiple available road paths, so that the cost of the transportation can be minimized to increase the profits of the logistics company and time of the transportation can be minimized, so that the customer satisfaction can be improved (Bruera  et al, 2008). The new monitoring system proposed makes use of the optimum A* algorithm along with heuristic function, so that the shortest path can be chosen among the multiple road paths.

The algorithm follows the logic in the following way. It visits the depth for the node selected and initially, it is considered as the best (Baum et al, 2013). If the visited node is found to be not a solution, it is returned to the destination node, through returning to the previous node, to find other and better promising node. The process recurs to visit the previous node until the destination node is found, which leads to the final solution (Botea, 2011). Google Static Maps service is used so that a map and relevant graph can be drawn. And Google Direction service is used as it helps to obtain an easy way to find the distance, in between the source or origin location to the destination location for transportation of the logistics company. Here, Google Geocoding service is used so that the coordinates of longitude and latitude of both origin and destination locations can be obtained.

The application and architecture helps to obtain the information about the routes available and the shortest path, with the real time data, including time, distance and so the cost of the transport. The application also shows the multiple available routes for both private and public transportation (Bauer et al, 2011). The proposed application can be accessible both through the mobile phone and website.

During the process, input is received by the server and the request along with the input is processed so that the requested information is then produced. Finally, the results are sent as output, in the form of the distance, time, cost and traffic conditions, in real time to the driver.

Conclusion

Logistics and transportation are the realistic and dynamic industries, which function, according to the real time data, about the distance, time and other data. And route planning plays vital and crucial role that determine the cost of transportation and so the profit of the logistics company. Hence, route algorithms became the primary concern, to ensure that the shortest path is selected for transporting the goods from source to destination. Monitoring system in logistics planning and optimization helps the drivers driving the vehicle with goods to select the shortest path. Shortest path calculation can be done by potential routing algorithms that are complex and comprehensive in nature.

Shortest paths are established through connecting the source and destination nodes, along with the intermediate paths. Satisfying path is finalized, after exploring all the sub-paths. Various algorithms for routing planning are explored and discussed in the project report. Different categorized algorithms, like construction, improvement and hybrid algorithms based on the heuristic algorithms are explored and discussed. Greedy algorithm is discussed with various forms and variations. Greedy algorithm is discussed, which is an improvement algorithm. Later, 2-opt and 3-opt algorithms are briefed about the logics, with which they are implemented. And greedy algorithm is further explored and discussed by the integration of the 2-opt algorithm, followed by 3-opt algorithm, with the greedy algorithm. After briefing of the algorithms of various logics and implementation, detailed exploration, discussion and analysis of the algorithms are presented starting with the Dijkstra algorithm, which is based on the graph theory. The algorithm combines the shortest path as well as the graph theory, so that the shortest graph is identified and decision process becomes easier for a driver, driving the transportation vehicle. The way this algorithm can be applicable to the logistics application is also explored and discussed and presented in the report. The next algorithm discussed is genetic algorithm that works on the simulation of the genetic solution process and natural elimination process. The algorithm basically works on exploring various and available range of options. So, on an overall three routing algorithms for route planning, called Greedy algorithm, Dijkstra algorithm and genetic algorithm are explored and discussed. Having understood these algorithms, these are compared with the advantages and disadvantages of them.

Having discussed the three algorithms, basically, a new and sophisticated algorithm, for Google Maps API and its potential are discussed with primary details. The Google Maps API is an integration of multiple Google Map related services and such application can be developed with a new and better algorithm. A new monitoring system is developed with basic and primary concept and logic that it can be implemented with. The architecture of the new monitoring system is developed, finally, with the basic input, process and output details.

References

Abraham, I, Delling, D, Fiat, A, Goldberg, A, V and Werneck, R, F (2013), Highway dimension and provably efficient shortest path algorithms. Technical Report MSRTR- 2013-91,  Microsoft Research.

Abraham, I, Delling, D, Goldberg, A, V and Werneck. R, F (2013), “Alternative routes in road networks”. ACM Journal of Experimental Algorithmics, 18(1):1–17.

Abraham, I, Delling, D, Goldberg, A, V Werneck. R, F, (2011), A hub-based labeling algorithm for shortest paths on road networks. In Proceedings of the 10th International Symposium on Experimental Algorithms (SEA’11), volume 6630 of Lecture Notes in Computer Science, pages 230–241. Springer.

Abraham, I, Delling, D, Goldberg, A, V, Werneck, R, F, (2012), Hierarchical hub labelings for shortest paths. In Proceedings of the 20th Annual European Symposium on Algorithms (ESA’12), volume 7501 of Lecture Notes in Computer Science, pages 24–35 Springer.

Abraham, I, Fiat, A, Goldberg, A, V and Werneck, R, F. (2010), Highway dimension, shortest paths, and provably efficient algorithms, In Proceedings of the 21st Annual ACM–SIAM Symposium on Discrete Algorithms (SODA’10), pages 782–793. SIAM.

Aifadopoulou, G, Ziliaskopoulos, A. and Chrisohoou, E, (2007), “Multiobjective optimum path algorithm for passenger pretrip planning in multimodal transportation networks”. Journal of the Transportation Research Board, 2032(1):26–34.

Akiba, T, Iwata, Y, Kawarabayashi, K. and Kawata, Y. (2014), Fast shortest-path distance queries on road networks by pruned highway labelling, In Proceedings of the 16^th Meeting on Algorithm Engineering and Experiments (ALENEX’14), pages 147–154. SIAM.

Akiba, T, Iwata, Y. and Yoshida, Y, (2013), Fast exact shortest-path distance queries on large networks by pruned landmark labelling, In Proceedings of the 2013 ACM SIGMOD International Conference on Management of Data (SIGMOD’13), pages 349–360. ACM Press.

Allulli, L, Italiano, G, F. Santaroni, F. (2014), Exploiting gps data in public transport journey planners. In Proceedings of the 13th International Symposium on Experimental Algorithms (SEA’14), volume 8504 of Lecture Notes in Computer Science. Springer.

Alshamrani, A., & Bahattab, A. (2015). A comparison between three SDLC models waterfall model, spiral model, and Incremental/Iterative model. International Journal of Computer Science Issues (IJCSI), 12(1), 106.

Antsfeld, L. and Walsh, T. (2012), Finding multi-criteria optimal paths in multi-modal public transportation networks using the transit algorithm. In Proceedings of the 19th ITS World Congress.

Arz, J, Luxen, D. and Sanders, P, (2013), Transit node routing reconsidered, In Proceedings of the 12th International Symposium on Experimental Algorithms (SEA’13), volume 7933 of Lecture Notes in Computer Science, pages 55–66. Springer.

Babenko, M, Goldberg, A, V, Gupta, A, Nagarajan, V, (2013), Algorithms for hub label optimization. In Proceedings of the 40th International Colloquium on Automata, Languages, and Programming (ICALP’13), volume 7965 of Lecture Notes in Computer Science, pages 69–80. Springer.

Bader, R, Dees, J, Geisberger, R. and Sanders, P, (2011), Alternative route graphs in road networks. In Proceedings of the 1st International ICST Conference on Theory and Practice of Algorithms in (Computer) Systems (TAPAS’11), volume 6595 of Lecture Notes in Computer Science, pages 21–32. Springer.

Barrett, C, Bisset, K, Holzer, M, Konjevod, G, Marathe, M, V. and  Wagner, D. (2008), Engineering label-constrained shortest-path algorithms. In Proceedings of the 4th International Conference on Algorithmic Aspects in Information and Management (AAIM’08), volume 5034 of Lecture Notes in Computer Science, pages 27–37. Springer.

Barrett, C, Bisset, K, Holzer, M, Konjevod, G, Marathe, M, V. and Wagner, D. (2009), Engineering label-constrained shortest-path algorithms. In The Shortest Path Problem: Ninth DIMACS Implementation Challenge, volume 74 of DIMACS Book, pages 309–319. American Mathematical Society.

Bast, H, Brodesser, M. and Storandt, S, (2013), Result diversity for multi-modal route planning. In Proceedings of the 13th Workshop on Algorithmic Approaches for Transportation Modeling, Optimization, and Systems (ATMOS’13), OpenAccess Series in Informatics (OASIcs),pages 123–136.

Bast, H, Carlsson, E, Eigenwillig, Geisberger, R, Harrelson, C, Raychev, V. and Viger, F. (2010), Fast routing in very large public transportation networks using transfer patterns. In Proceedings of the 18th Annual European Symposium on Algorithms (ESA’10), volume 6346 of Lecture Notes in Computer Science, pages 290–301. Springer.

Bast, H, Funke, S, Matijevic, D, Sanders, P. and Schultes, D. (2007), In transit to constant shortest-path queries in road networks, In Proceedings of the 9th Workshop on Algorithm Engineering and Experiments (ALENEX’07), pages 46–59. SIAM.

Bast, H, Funke, S, Sanders, P. and Schultes,D. (2007), Fast routing in road networks with transit nodes, Science, 316(5824):566.

Bast, H, Funke, S. and Matijevic,D. (2009), Ultrafast shortest-path queries via transit nodes. In The Shortest Path Problem: Ninth DIMACS Implementation Challenge, volume 74 of DIMACS Book, pages 175–192. American Mathematical Society.

Bast, H, Sternisko, J. and Storandt, S, (2013), Delay-robustness of transfer patterns in public transportation route planning, In Proceedings of the 13th Workshop on Algorithmic Approaches for Transportation Modeling, Optimization, and Systems (ATMOS’13), OpenAccess Series in Informatics (OASIcs), pages 42–54.

Bast, H. and Storandt, S (2014), Frequency-based search for public transit, In Proceedings of the 22nd ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems. ACM Press 2014.

Bast, H. and Storandt, S. (2014), Flow-based guidebook routing. In Proceedings of the 16^th Meeting on Algorithm Engineering and Experiments (ALENEX’14), pages 155–165. SIAM.

Bast,H, (2009), Car or public transport – two worlds. In Efficient Algorithms, volume 5760 of Lecture Notes in Computer Science, pages 355–367. Springer.

Batz, G, V, Geisberger, R, Luxen, D, Sanders, P. and Zubkov, R. (2012), Efficient route compression for hybrid route planning. In Proceedings of the 1st Mediterranean Conference on Algorithms. Springer.

Batz, G, V, Geisberger, R, Sanders, P. and Vetter, C. (2013), “Minimum timedependent travel times with contraction hierarchies”, ACM Journal of Experimental Algorithmics, 18(1.4):1–43.

Batz, G, V. and Sanders, P, (2012), Time-dependent route planning with generalized objective functions. In Proceedings of the 20th Annual European Symposium on Algorithms (ESA’12), volume 7501 of Lecture Notes in Computer Science. Springer.

Bauer, A, (2012), Multimodal profile queries, Bachelor thesis, Karlsruhe Institute of Technology, 2012.

Bauer, R, Baum, M, Rutter, I. and Wagner, D, (2013), “On the complexity of partitioning graphs for arc-flags”, Journal of Graph Algorithms and Applications, 17(3):265–299.

Bauer, R, Columbus, T, Katz, B, Krug, M. and Wagner, D, (2010), Preprocessing speed-up techniques is hard. In Proceedings of the 7th Conference on Algorithms and Complexity (CIAC’10), volume 6078 of Lecture Notes in Computer Science, pages 359–370. Springer.

Bauer, R, Columbus, T, Rutter, I. and Wagner, D, (2013), Search-space size in contraction hierarchies. In Proceedings of the 40th International Colloquium on Automata, Languages, and Programming (ICALP’13), volume 7965 of Lecture Notes in Computer Science, pages 93–104. Springer.

Bauer, R, D’Angelo, G, Delling, D, Schumm, A. and Wagner, D. (2012), “The shortcut problem – complexity and algorithms”. Journal of Graph Algorithms and Applications, 16(2):447–481.

Bauer, R, Delling, D, Sanders, P, Schieferdecker, D, Schultes, D. and Wagner, D. (2010). “Combining hierarchical and goal-directed speed-up techniques for Dijkstra’s algorithm”. ACM Journal of Experimental Algorithmics, 15(2.3):1–31. Special Section devoted to WEA’08.

Bauer, R, Delling, D. and Wagner, D. (2011). Experimental study on speed-up techniques for timetable information systems. Networks, 57(1):38–52.

Bauer, R, Krug, M, Meinert, S. and Wagner, D. (2010). Synthetic road networks. In Proceedings of the 6th International Conference on Algorithmic Aspects in Information and Management (AAIM’10), volume 6124 of Lecture Notes in Computer Science, pages 46–57. Springer.

Bauer, R. and Delling, D, (2009), “SHARC: Fast and robust unidirectional routing. ACM” Journal of Experimental Algorithmics, 14(2.4):1–29. Special Section on Selected Papers from ALENEX 2008.

Baum, M, Dibbelt, J, Hübschle-Schneider, L, Pajor, T. and Wagner, D. (2014). Speed-consumption tradeoff for electric vehicle route planning. In Proceedings of the 14th Workshop on Algorithmic Approaches for Transportation Modeling, Optimization, and Systems (ATMOS’14), OpenAccess Series in Informatics (OASIcs), pages 138–151.

Baum, M, Dibbelt, J, Pajor, M. and Wagner, D. (2013). Energy-optimal routes for electric vehicles. In Proceedings of the 21st ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, pages 54–63. ACM Press.

Berger, A, Delling, D, Gebhardt, A. and Müller–Hannemann, M, (2009). Accelerating time-dependent multi-criteria timetable information is harder than expected. In Proceedings of the 9th Workshop on Algorithmic Approaches for Transportation Modeling, Optimization, and Systems (ATMOS’09), OpenAccess Series in Informatics (OASIcs).

Berger, A, Gebhardt, A, Müller–Hannemann, M. and Ostrowski, M. (2011). Stochastic delay prediction in large train networks. In Proceedings of the 11th Workshop on Algorithmic Approaches for Transportation Modeling, Optimization, and Systems (ATMOS’11), volume 20 of OpenAccess Series in Informatics (OASIcs), pages 100–111.

Berger, A, Grimmer, M. and Müller–Hannemann, M. (2010). Fully dynamic speed-up techniques for multi-criteria shortest path searches in time-dependent networks. In Proceedings of the 9th International Symposium on Experimental Algorithms (SEA’10), volume 6049 of Lecture Notes in Computer Science, pages 35–46. Springer.

Bielli, M, Boulmakoul, A. and Mouncif, H. (2006). “Object modeling and path computation for multimodal travel systems”. European Journal of Operational Research, 175(3):1705–1730.

Böhmová, K, Mihalák, M, Pröger, T, Šrámek, R. and Widmayer, P. (2013). Robust routing in urban public transportation: How to find reliable journeys based o past observations. In Proceedings of the 13th Workshop on Algorithmic Approaches for Transportation Modeling, Optimization, and Systems (ATMOS’13), OpenAccess Series in Informatics (OASIcs), pages 27–41, 2013.

Botea, A. (2011). Ultra-fast optimal pathfinding without runtime search. In Proceedings of the Seventh AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE’11), pages 122–127. AAAI Press.

Botea, A. and Harabor, D. (2013).  Path planning with compressed all-pairs shortest paths data. In Proceedings of the 23rd International Conference on Automated Planning and Scheduling. AAAI Press.

Brandes, U, Schulz, F, Wagner, D. and Willhalm, T. (2001). Travel planning with self-made maps. In Proceedings of the 3rd International Workshop on Algorithm Engineering and Experiments (ALENEX’01), volume 2153 of Lecture Notes in Computer Science, pages 132–144. Springer.

Brandes, U. and Erlebach, T. (2005).  Network Analysis: Methodological Foundations, volume 3418 of Lecture Notes in Computer Science. Springer.

Brodal,G. and Jacob, R. (2004). Time-dependent networks as models to achieve fast exact time-table queries. In Proceedings of the 3rd Workshop on Algorithmic Methods and Models for Optimization of Railways (ATMOS’03), volume 92 of Electronic Notes in Theoretical Computer Science, pages 3–15.

Bruera, F, Cicerone, S, D’Angelo, G, Stefano, G. and Frigioni, D. (2008). Dynamic multi-level overlay graphs for shortest paths. Mathematics in Computer Science, 1(4):709–736.

Brunel, E, Delling, D, Gemsa, A. and Wagner, A. (2010). Space-efficient SHARCrouting. In Proceedings of the 9th International Symposium on Experimental Algorithms (SEA’10), volume 6049 of Lecture Notes in Computer Science, pages 47–58. Springer.

Cattaruzza, D., Absi, N., Feillet, D., & González-Feliu, J. (2017). Vehicle routing problems for city logistics. EURO Journal on Transportation and Logistics, 6(1), 51-79. 10.1007/s13676-014-0074-0

Cohen, E, Halperin, E, Kaplan, H. and Zwick, U. (2003). “Reachability and distance queries via 2-hop labels”. SIAM Journal on Computing, 32(5):1338–1355.

Crainic, T, G, Florian, G, Noriega, Y, (2007), Computing shortest paths with logistics constraints, CIRRELT Noriega

Crainic, T. G., Ricciardi, N., & Storchi, G. (2009). Models for evaluating and planning city logistics systems. Transportation Science, 43(4), 432-454. 10.1287/trsc.1090.0279

D’Angelo, G, D’Emidio, M, Frigioni, D. and Vitale, C, (2012). Fully dynamic maintenance of arc-flags in road networks. In Proceedings of the 11th International Symposium on Experimental Algorithms (SEA’12), volume 7276 of Lecture Notes in Computer Science, pages 135–147. Springer.

Dehne, F, Omran, M, T. and Sack,J. (2012). Shortest paths in time-dependent fifo networks. Algorithmica, 62:416–435.

Delling, D, Dibbelt, J, Pajor, T, Wagner, D. and Werneck, R, F. (2013). Computing multimodal journeys in practice. In Proceedings of the 12th International Symposium on Experimental Algorithms (SEA’13), volume 7933 of Lecture Notes in Computer Science, pages 260–271. Springer.

Delling, D, Giannakopoulou, K, DorotheaWagner. and Zaroliagis, C. (2008). Timetable information updating in case of delays: Modeling issues. Technical Report 133, Arrival Technical Report.

Delling, D. (2011). Time-dependent SHARC-routing. Algorithmica, 60(1):60–94 2011.

Goldberg, D.E. (1989), Genetic Algorithms in Search, Optimization and Machine Learning, Addison Welsey Publishing Company.

Golden, B., (1976), Shortest-Path Algorithms – A comparison, Operations Research

Hao, H, N, Suling, W, (2009), “Modeling and solving for locations of multi-level logistics network”, Journal of Shanghai Maritime University.,(4).

Hao, H, Suling, W, (2009), “Modeling and solving for locations of multi-level logistics network”, Journal of Shanghai Maritime University

Hui, W, Chuan-xiang, R, Chang-chang, Y and Xin-ga, H, (2009), “Study on optimization of logistics distribution route based on niche genetic algorithm”, Journal of Computer Applications

Nabil Mohammed Ali Munassar, & Govardhan, A. (2010). A comparison between five models of software engineering. International Journal of Computer Science Issues (IJCSI), 7(5), 94.

Raichaudhuri, A, Jain, A, (2010), ‘Genetic algorithm based logistics route planning’, International journal of nnovation, management and technology, Vol, 1, No. 2.

Xiao-nian, H, E and Xiao-liang, X, (2009), Capacitated logistic distribution vehicle routing optimization. Computer Engineering and Applications

Xiao-Yan, L, Yan-Li, C, (N.Y), Application of Dijkstra algorithm in logistics distribution lines, Proceedings of the third international symposium in computer science and computational technology (ISCSCT ’10), China: Henan Polytechnic University

Yang, L, E and Jian-ya, G,(1999), An efficient implementation of shortest path algorithm based on Dijkstra algorithm, Geomatics and Information Science of Wuhan University.

Zhang, Y., Shi, L., Chen, J., & Li, X. (2017). Analysis of an automated vehicle routing problem in logistics considering path interruption. Journal of Advanced Transportation, 201710.1155/2017/1624328